Evaluation of the efficacy of FDG-PET/CT in assessing the effects of BMMSC treatment in aged rhesus monkeys

Abstract

To evaluate the therapeutic efficacy of BMMSCs in aged rhesus monkeys, histological texture features were analyzed using ¹⁸F-FDG-PET/CT imaging. The study involved three groups: five juvenile rhesus monkeys (mean age 2 years, weight 4.6 ± 0.49 kg), five young adult rhesus monkeys (mean age 6.2 ± 0.40 years, weight 3.6 ± 0.80 kg), and six aged rhesus monkeys (mean age 23.3 ± 1.37 years, weight 7.0 ± 2.77 kg). The aged group underwent ¹⁸F -FDG-PET/CT imaging at baseline, and again at 3 and 6 months following BMMSC administration. Image texture features were segmented, extracted, and analyzed using LIFEx software, with a total of 93 high-throughput PET texture features evaluated across the adolescent, young adult, and aged groups, as well as at the 3- and 6-month post-treatment time points.Significant differences (P < 0.05) in 82 texture features were observed in multiple organs, including the brain, bilateral retroocular fat, thyroid, thymus, lung, aorta, gastric sinusoids, liver, kidneys, and L4 vertebrae, across the three age groups. Comparison between the younger and older groups revealed significant differences (P < 0.05) in 31 texture features within the brain, bilateral retroocular fat, thyroid, lung, gastric sinus, left kidney, and L4 vertebrae.ROC curve analysis of the PET texture features from the thyroid, lung, gastric sinus, left kidney, and L4 vertebrae showed AUC values greater than 0.7, indicating good diagnostic performance. In aged rhesus monkeys treated with BMMSCs, PET texture features in the brain, bilateral retroocular fat, left lung, gastric sinus, and L4 vertebrae recovered to levels comparable to those of the young group at both the 3- and 6-month post-treatment time points. Additionally, histological examination via HE staining of the lung, liver, kidney, stomach, and thymus in the aged group at 6 months post-treatment revealed significant improvements in the structural integrity of these organs.These findings suggest that ¹⁸F -FDG PET texture analysis is a valuable tool for assessing the restoration of physiological function in aged rhesus monkeys following BMMSC treatment.

Introduction

Aging is a biological process that commences after sexual maturity, characterized by a gradual decline in molecular integrity, reduced tissue and cellular function, and a deterioration in the performance of multiple organs. This process increases the susceptibility to age-related diseases and ultimately contributes to morbidity and mortality [1, 2]. According to the 2015 United Nations Global Report on Population Ageing, life expectancy is rising globally. By 2050, the number of people aged 60 and over is expected to double, reaching nearly 2.1 billion [3]. As aging progresses, systemic organ failure becomes evident, manifesting as cognitive and memory decline, reduced visual function, impaired thyroid activity, thymic atrophy, respiratory and circulatory dysfunction, increased digestive and urinary burdens, and degenerative changes in the vertebrae. Slowing the aging process, restoring organ function, and improving the quality of life in the elderly have become critical and pressing objectives.

Stem cells possess the remarkable ability to differentiate into various cell types within the body and can proliferate indefinitely [4]. Mesenchymal stem cells (MSCs) are a subpopulation of multipotent stem cells characterized by low immunogenicity, self-renewal capacity, and the potential for multidirectional differentiation. Through paracrine signaling, MSCs secrete bioactive molecules with immunomodulatory, anti-apoptotic, pro-angiogenic, anti-fibrotic, and chemotactic properties, which have been utilized to alleviate a variety of diseases [5]. Bone marrow-derived mesenchymal stem cells (BMMSCs) not only maintain low immunogenicity but also exhibit potent immunomodulatory capabilities and secrete a variety of cytokines both in vitro and in vivo. These cells contribute to hematopoiesis and support tissue and organ regeneration by promoting immunomodulation and chemotaxis at sites of damage [6]. As stem cells with multipotent differentiation potential, BMMSCs can differentiate into osteoblasts, chondrocytes, adipocytes, and other cell types, while also secreting a variety of cytokines and bioactive factors [7]. BMMSCs have shown remarkable potential for treating age-related diseases and promoting tissue repair due to their unique properties, making them an important biotherapeutic approach for combating aging. Given the physiological and genomic similarities between rhesus monkeys and humans (approximately 93%), studies on the therapeutic effects of MSCs in aged rhesus monkeys offer valuable insights into potential interventions to mitigate aging in humans [8].Our previous studies have demonstrated that improved BMMSCs enhance the structure and function of organs such as the lungs, thymus, and ovaries in aged rhesus monkeys, as evidenced by transcriptome sequencing, protein-protein interaction (PPI) network analysis, and cytomorphological examination [9, 10]. However, due to ethical concerns associated with in vivo tissue extraction, the application of these methods in humans remains limited, particularly when validating cytological, proteomic, and genomic findings. To overcome this limitation, we utilized epigenetic clocks and age-sensitive in-depth phenotyping to provide a multidimensional perspective on aging and its effects on various tissues, organs, and systems, thereby opening new avenues for novel approaches to studying aging-related phenotypic outcomes [11–13].

Imaging histology is a noninvasive examination technique that extracts a large number of high-throughput imaging features from medical images for analysis, offering advantages such as safety, convenience, speed, and reproducibility. Age-sensitive deep phenotyping, in combination with imaging histology, provides novel insights for assessing human efficacy [14, 15]. Positron Emission Tomography (PET) is a non-invasive imaging technique that utilizes molecular tracing, offering high sensitivity, high specificity, and whole-body imaging capabilities. By detecting changes in the glucose metabolic rate of local tissues, PET imaging reflects alterations in certain physiological functions in vivo, revealing both anatomical and functional metabolic changes in tissues. PET image texture characterization is a well-established image processing technique that has demonstrated predictive potential in various malignant tumors, and it can assess tumor conditions by quantifying tumor heterogeneity. Texture or heterogeneity analysis via PET imageomics provides more comprehensive information about whole-body, multi-organ-related phenotypes than simple measurements based on SUV alone. To date, no study has reported the use of PET texture in assessing the efficacy of systemic multi-organ stem cell therapy. The aim of this study was to utilize ¹⁸F-FDG-PET/CT imaging to assess the efficacy of BMMSCs in treating aged rhesus monkeys by analyzing image texture features.

Materials and methods

All animal experiments in this study were conducted in accordance with the ARRIVE guidelines (version 2.0) to ensure rigorous experimental design, implementation, and transparent reporting of results.

Subject of study

In this experiment, female rhesus monkeys housed in captivity were selected as experimental subjects and underwent physical examinations and behavioral assessments conducted by professional technicians. Sixteen healthy rhesus monkeys with no history of brain diseases (including traumatic brain injury, brain tumors, mental illness, or brain surgery) were screened and grouped according to age. The groups consisted of 5 monkeys in the juvenile group (mean age 2 years, weight 4.6 ± 0.49 kg), 5 monkeys in the young group (mean age 6.2 ± 0.40 years, weight 3.6 ± 0.80 kg), and 6 monkeys in the elderly group (mean age 23.3 ± 1.37 years, weight 7.0 ± 2.77 kg).

Culture and transplantation of BMMSCs

This study used healthy young female rhesus macaques with an average age of 1 year to extract bone marrow. After extraction, culture, passage, viability testing, and identification, P4 generation BMMSCs (with fusion degree passage rate > 80%) were obtained. Sixteen rhesus macaques were kept under normal conditions for one week without abnormal changes, and were then grouped by age: juvenile group (n = 5), young group (n = 5), and elderly group (n = 6). The elderly treatment group received intravenous transplantation of 1 × 10⁷ BMMSCs/kg via femoral vein, with one infusion every other day for a total of three infusions, while being regularly fed. This study has been approved by the Ethics Committee of the 920th Hospital of the Joint Logistics Support Force of the People’s Liberation Army (approval number: Lunshen 2019-003(Ke)-01). Animal care was carried out according to the guidelines set by the China National Accreditation Service for Conformity Assessment (CNASLA0001). All experimental animals were housed at the Kunming Institute of Zoology, Chinese Academy of Medical Sciences [SYXK(Dian)K2017-0008], and all experimental procedures followed the “Policy on Humane Care and Use of Laboratory Animals” of the U.S. Public Health Service.

¹⁸F-FDG-PET image acquisition

Prepare before imaging

The glucose metabolism imaging agent ¹⁸F-FDG was produced by GE, USA, and synthesized using the MINItrace Kirin cyclotron and the Tracerlab FDG synthesizer, achieving a radiochemical purity of > 99%. Rhesus monkeys were fasted for 6 h prior to imaging. Before the procedure, the monkeys were anesthetized with 3% pentobarbital sodium (1 mL/kg body mass), and fasting blood glucose levels were measured to ensure they were < 8 mmol/L. Subsequently, ¹⁸F-FDG was intravenously injected at a dose of 0.3 mCi/kg body mass. The monkeys were then placed in a quiet, darkened room and allowed to rest for 40 min after injection.

Image scanning method

Whole-body images of rhesus monkeys were acquired using an ¹⁸ F-FDG PET/CT scanner (DiscoveryTM PET/CT Elite, GE Healthcare, USA), with their vital signs closely monitored throughout the procedure. The monkeys were positioned supine on the scanning bed, ensuring that the canthus line was perpendicular to the bed, the trunk was parallel to the bed, and the upper limbs were naturally placed at the sides of the body. First, a conventional spiral CT scan of the head was performed, covering the area from the top of the skull to the middle of the neck, followed by PET imaging. The monkey number, weight, age, and injection dose were recorded. The acquisition parameters were set as follows: tube voltage 120 kV, tube current 260 mA, pitch 0.561, rotational speed 0.5 s/rotation, slice thickness 3.75 mm, slice spacing 3.75 mm, matrix 512 × 512, and field of view (FOV) 50 cm × 50 cm. PET acquisition time was 2.5 min per bed position. After image acquisition was completed, the monkeys were returned to their cages to recover naturally. Imaging was performed for the juvenile, young, and old age groups, as well as for the old age group at 3 and 6 months post-treatment, respectively. The imaging procedure is shown in Fig. 1.

Fig. 1.

PET/CT whole-body images. (A-E) Whole-body images of rhesus monkeys in the juvenile, young, old, 3-month post-treatment, and 6-month post-treatment groups, with separate scans of the head and the rest of the body

Image processing, segmentation, and texture extraction

PET image preprocessing

The PET and CT images were transferred to the AW4.6 workstation, where CT data attenuation correction was applied to obtain fused PET/CT images. After scanning, a senior physician conducted an image evaluation, checking for proper body positioning, absence of motion artifacts, clear image resolution, no abnormal increases or decreases in FDG radioactivity distribution, and good alignment between the PET and CT images. This process ensured that the image quality was reliable, the signal-to-noise ratio was adequate, and any images that did not meet the required standards were re-scanned or discarded.

Image segmentation

All full-body PET and CT images of the rhesus macaques were extracted. First, the images were exported in DICOM format using the GE AW4.6 workstation. Then, they were imported into the open-source medical image processing and navigation software LIFEx 7.1.0 for annotation of the target organs [16]. Initially, one physician used the software’s annotation tool to manually outline the regions of interest (ROIs) on each PET/CT image layer using a blinded method, with the CT images assisting in localization. Then, another senior physician verified and corrected the outlined results. This process is illustrated in Fig. 2.

Fig. 2.

Target organ maps of PET/CT images in the juvenile, young, geriatric, 3-month post-treatment, and 6-month post-treatment groups. The regions of interest outlined in the juvenile, young, geriatric, 3-month post-treatment, and 6-month post-treatment groups include, in order: brain, right and left retroocular fat, thyroid, thymus, aorta, left ventricle, both lungs, liver, stomach, pancreas, both kidneys, spleen, and L4 vertebrae

Texture extraction

Ninety-three high-throughput PET radiographic features, including shape features, first-order features, and second-order features, were automatically calculated and extracted using the LIFEx software program, as detailed in the Supplementary Material [17], All analyses were performed by trained technicians in a blinded manner to minimize researcher bias and ensure the objectivity and reliability of the results.

Statistical analysis

SPSS 21 software was used to test the normality of the quantitative data. Data conforming to a normal distribution are presented as , while data not conforming to a normal distribution are presented as median (Q1, Q3). Differences between multiple groups were compared using one-way ANOVA or the Kruskal-Wallis test, and pairwise differences were analyzed using the independent samples t-test, Mann-Whitney U-test, and chi-square test (χ²). Statistical analysis was performed on three groups of rhesus monkeys (juvenile, young, and old) to identify age-sensitive PET texture features. Differential comparisons between the young and old groups were subsequently conducted to analyze texture differences. The classification ability of the textures was assessed using ROC curve analysis, and the classification threshold was determined when the Youden index reached its maximum value. The efficacy of BMMSCs in aged rhesus monkeys was evaluated by using young rhesus monkeys as the control group. A p-value of < 0.05 was considered statistically significant.

Collection of organ histological samples and hematoxylin and Eosin (H&E) staining

Six months after BMMSC treatment, the macaques were anesthetized and euthanized with 3% sodium pentobarbital. The lungs, liver, kidneys, stomach, and thymus glands were rapidly excised via an open abdominal procedure, cleaned, weighed, and photographed for documentation. Each organ was then cut into approximately 1 cm³ tissue blocks, which were further sectioned into 1 mm³ pieces. These pieces were placed in 4% paraformaldehyde fixative at 4 °C for 24 h, followed by routine dehydration, clearing, and paraffin embedding. Paraffin blocks were sectioned to a thickness of approximately 4 μm, and HE staining was performed. The sections were sequentially dewaxed, rehydrated, stained, dehydrated, cleared, and mounted with neutral gum. Finally, tissue structures were observed under a light microscope, photographed, and recorded for pathological analysis.

Results

Texture extraction

Ninety-three high-throughput texture features were extracted from the juvenile, young, old, 3-month post-treatment, and 6-month post-treatment groups, including 5 shape features, 13 first-order features, and 31 s-order features.

Differential analysis of PET texture between the juvenile, young, and elderly rhesus macaque groups

In the brain, 2 texture features were statistically different; in the right retroocular fat, 15 texture features were statistically different; in the left retroocular fat, 5 texture features were statistically different; in the thyroid, 3 texture features were statistically different; in the thymus, 4 texture features were statistically different; in the aorta, 5 texture features were statistically different; in the left ventricle, no texture features were statistically different; in the right lung, 10 texture features were statistically different; in the left lung, 16 texture features were statistically different; in the stomach, 6 texture features were statistically different; in the liver, 4 texture features were statistically different; in the right kidney, 1 texture feature was statistically different; in the left kidney, 4 texture features were statistically different; in the L4 vertebra, 1 texture feature was statistically different. PET textures in the left ventricle, pancreas, and spleen were not statistically different between young and aged rhesus monkeys. These results are summarized in Table 1.

Table 1.

Differential analysis of PET texture among the juvenile, young, and elderly rhesus macaque groups

Organ	Radiomics Feature(RF)	P_value	Analysis
Brain	SHAPE_Sphericity[onlyFor3DROI]	0.037	One-way analysis
	GLZLM_ZLNU	0.024	KW
R-FAT	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	0.008	KW
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	0.005	KW
	SHAPE_Volume(mL)	0.001	KW
	SHAPE_Volume(vx)	0.001	KW
	SHAPE_Sphericity[onlyFor3DROI])	0.006	KW
	SHAPE_Compacity[onlyFor3DROI]	0.009	KW
	GLRLM_RLNU	0.003	KW
	NGLDM_Busyness	0.047	KW
	GLZLM_LZE	0.026	KW
	GLZLM_LZLGE	0.028	KW
	GLZLM_LZHGE	0.021	KW
	GLZLM_GLNU	0.006	KW
	NGLDM_Coarseness	0	One-way analysis
	SHAPE_Surface(mm2)[onlyFor3DROI]	0.036	One-way analysis
	GLCM_Correlation	0	One-way analysis
F-FAT	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	0.005	KW
	DISCRETIZED_SUVbwmin	0.021	KW
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	0.004	KW
	SHAPE_Volume(mL)	0	One-way analysis
	SHAPE_Volume(vx)	0	One-way analysis
Thyriod	CONVENTIONAL_SUVbwQ2	0.019	KW
	DISCRETIZED_SUVbwQ3	0.025	KW
	DISCRETIZED_SUVbwpeakSphere1mL	0.029	KW
R-Lung	SHAPE_Volume(mL)	0.003	One-way analysis
	SHAPE_Volume(vx)	0.003	One-way analysis
	SHAPE_Surface(mm2)[onlyFor3DROI]	0	One-way analysis
	GLCM_Correlation	0.037	One-way analysis
	GLRLM_GLNU	0.004	One-way analysis
	GLZLM_LZHGE	0.014	One-way analysis
	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	0.004	KW
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	0.003	KW
	GLRLM_RLNU	0.022	KW
	GLRLM_RP	0.013	KW
L-Lung	SHAPE_Surface(mm2)[onlyFor3DROI]	0	One-way analysis
	SHAPE_Compacity[onlyFor3DROI]	0	One-way analysis
	GLRLM_SRE	0.034	One-way analysis
	NGLDM_Busyness	0.016	One-way analysis
	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	0.005	KW
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	0.002	KW
	SHAPE_Volume(mL)	0.001	KW
	SHAPE_Volume(vx)	0.001	KW
	GLRLM_GLNU	0.007	KW
	GLRLM_RLNU	0.007	KW
	NGLDM_Coarseness	0.001	KW
	GLZLM_LZHGE	0.041	KW
	GLZLM_GLNU	0.027	KW
	DISCRETIZED_SUVbwQ3	0.034	KW
	GLRLM_LRE	0.036	KW
	GLRLM_RP	0.034	KW
Thymus	SHAPE_Volume(mL)	0	One-way analysis
	SHAPE_Volume(vx)	0	One-way analysis
	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	0.005	KW
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	0.005	KW
Artery	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	0.03	One-way analysis
	DISCRETIZED_SUVbwmean	0.005	One-way analysis
	DISCRETIZED_SUVbwstd	0.005	One-way analysis
	DISCRETIZED_HISTO_Entropy_log10	0.043	KW
	DISCRETIZED_HISTO_Entropy_log2	0.043	KW
Liver	DISCRETIZED_SUVbwSkewness	0.016	KW
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	0.044	KW
	GLZLM_LZHGE	0.018	KW
	SHAPE_Surface(mm2)[onlyFor3DROI]	0.029	One-way analysis
Stomach	CONVENTIONAL_SUVbwKurtosis	0.045	One-way analysis
	CONVENTIONAL_SUVbwExcessKurtosis	0.045	One-way analysis
	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	0.039	KW
	DISCRETIZED_SUVbwmin	0.033	KW
	SHAPE_Volume(mL)	0.018	KW
	SHAPE_Volume(vx)	0.018	KW
R-kidney	SHAPE_Compacity[onlyFor3DROI]	0.022	One-way analysis
L-kidney	GLCM_Correlation	0.032	One-way analysis
	SHAPE_Compacity[onlyFor3DROI]	0.026	KW
	GLRLM_RLNU	0.033	KW
	NGLDM_Coarseness	0.049	KW
L4	SHAPE_Volume(mL)	0.037	One-way analysis
	SHAPE_Volume(vx)	0.037	One-way analysis
	SHAPE_Sphericity[onlyFor3DROI])	0.008	One-way analysis
	SHAPE_Surface(mm2)[onlyFor3DROI]	0.01	One-way analysis
	GLRLM_RLNU	0.034	One-way analysis
	GLZLM_GLNU	0.003	One-way analysis
	GLZLM_ZLNU	0.037	KW
Left Ventricular		> 0.05
Pancreas		> 0.05
Spleen		> 0.05

Differential analysis of PET texture between young and elderly rhesus macaques

Statistical analyses were conducted on textures that significantly differed between the three groups, and PET texture differences were compared between the young and old age groups. The results showed that in the brain, one texture was significantly different, with the older group having higher values than the younger group, and the younger group having higher values than the juvenile group. In the right retroocular fat, 10 textures differed significantly, with most textures showing higher values in the older group compared to the younger group, and the younger group having higher values than the juvenile group. However, for the SHAPE_Sphericity[onlyFor3DROI] and NGLDM_Coarseness textures, the juvenile group had higher values than the younger group, and the younger group had higher values than the older group. In the left retroocular fat, 4 textures were significantly different, with the older group having higher values than the younger group, and the younger group having higher values than the juvenile group. In the thyroid, one texture showed a significant difference, with the older group having higher values than the younger group, and the juvenile group having higher values than the younger group. In the right lung, 3 textures differed statistically, with the older group having higher values than the younger group, and the younger group having higher values than the juvenile group. In the left lung, 7 textures showed significant differences, with the older group having higher values than the younger group, and the younger group having higher values than the juvenile group. In the stomach, 3 textures showed significant differences, with the older group having higher values than the juvenile group, and the juvenile group having higher values than the younger group. In the left kidney, 1 texture showed a significant difference, with the older group having higher values than the younger group, and the younger group having higher values than the juvenile group. In the L4 vertebra, 1 texture showed a significant difference, with the older group having higher values than the younger group, and the younger group having higher values than the juvenile group. Conversely, PET textures in the thymus, aortic arch, left ventricle, liver, pancreas, right kidney, and spleen did not show significant differences between young and aged rhesus monkeys. Related data and results can be found in Table 2; Fig. 3.

Table 2.

Differential analysis of PET texture between the young and elderly rhesus macaque groups

Organ	Radiomics Feature(RF)	Young	Elderly	P_value	Analysis
Brain	SHAPE_Sphericity[onlyFor3DROI]	0.707 ± 0.010	0.735 ± 0.012	0.006	LSD
R-Fat	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	1.025 ± 0.613	2.408 ± 0.817	0.018	LSD
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	3.217(2.032,6.155)	9.213(7.095,10.911)	0.018	M-W U
	SHAPE_Volume(mL)	1.325(0.97,1.669)	2.778(2.387,3.245)	0.006	M-W U
	SHAPE_Volume(vx)	166(121,209)	348(299,407)	0.006	M-W U
	SHAPE_Sphericity[onlyFor3DROI])	0.46(0.458,0.549)	0.415(0.402,0.442)	0.028	M-W U
	SHAPE_Surface(mm2)[onlyFor3DROI]	1194 ± 317	2346 ± 468	0	LSD
	GLRLM_RLNU	85(67,102)	170(128,200)	0.018	M-W U
	NGLDM_Coarseness	0.053 ± 0.014	0.028 ± 0.007	0.01	LSD
	GLZLM_GLNU	3.364(2.929,4)	4.776(4.498,5.058)	0.018	M-W U
	GLCM_Correlation	0.431 ± 0.143	0.596 ± 0.141	0.023	LSD
L-Fat	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	1.066(0.607,1.643)	2.462(1.927,2.927)	0.018	M-W U
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	4.064(2.431,6.083)	9.349(7.303,11.129)	0.018	M-W U
	SHAPE_Volume(mL)	1.289 ± 0.322	3.048 ± 0.832	0.006	LSD
	SHAPE_Volume(vx)	161 ± 40	382 ± 104	0.045	LSD
Thyroid	CONVENTIONAL_SUVbwQ2	0.588(0.0.535,0.673)	0.901(0.683,0.995)	0.018	M-W U
R-Lung	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	50(36,59)	69(66,91)	0.006	M-W U
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	229(155,258)	294(289,368)	0.006	M-W U
	SHAPE_Surface(mm2)[onlyFor3DROI]	21,630 ± 6141	27,464 ± 2944	0.023	LSD
L-Lung	SHAPE_Volume(mL)	109(106,139)	172(155,186)	0.017	M-W U
	SHAPE_Volume(vx)	2501(2443,3196)	3966(3564,4283)	0.017	M-W U
	SHAPE_Surface(mm2)[onlyFor3DROI]	21,456 ± 2862	27,197 ± 2295	0.011	LSD
	SHAPE_Compacity[onlyFor3DROI]	5.55 ± 0.532	6.305 ± 0.432	0.028	LSD
	GLRLM_GLNU	436(390,481)	580(484,622)	0.018	M-W U
	GLZLM_LZHGE	85,839(51603,211465)	345,988(176262,468263)	0.045	M-W U
	NGLDM_Busyness	14.147 ± 7.690	24.213 ± 10.298	0.041	LSD
Stomach	CONVENTIONAL_SUVbwKurtosis	2.224 ± 0.636	3.263 ± 0.769	0.041	LSD
	CONVENTIONAL_SUVbwExcessKurtosis	-0.776 ± 0.636	0.263 ± 0.769	0.041	LSD
	SHAPE_Volume(mL)	0.391(0.283,2.021)	2.347(2.032,3.673)	0.045	M-W U
L-Kidney	GLRLM_RLNU	111(65,167)	194(157,362)	0.045	M-W U
L4	GLZLM_GLNU	2.137 ± 0.595	4.056 ± 1.135	0.006	LSD
Thymus				> 0.05
Aorta				> 0.05
Left Ventricular				> 0.05
Liver				> 0.05
Pancreas				> 0.05
R-Kidney				> 0.05
Spleen				> 0.05

Fig. 3.

Box Plot of PET texture for the Juvenile, Young, and Elderly Rhesus Macaque group. In this experiment, the PET textures that differed among the three groups were further analyzed statistically by comparing the young group with the aged group. A. Brain: Box plots showing the differences in texture between the young and aged rhesus monkeys across different age groups; B-E. Posterior Ocular Fat: Box plots illustrating the differences in texture between the young and aged rhesus monkeys across different age groups; F. Thyroid: Box plot of the differences in texture between the young and aged rhesus monkeys across different age groups; G-I. Lungs: Box plots comparing the texture differences between young and aged rhesus monkeys across different age groups; J. Gastric Sinus: Box plots showing the differences in texture between young and aged rhesus monkeys across different age groups; K. Left Kidney: Box plot illustrating the texture differences between the young and aged rhesus monkeys across different age groups; L. L4: Box plots showing the differential texture results between young and old rhesus monkeys, and the box plots for the different age groups

Evaluation of the effectiveness of PET texture

The SPSS ROC analysis of the textures that differed between the young and old age groups revealed the following results: In the brain, the texture HAPE_Sphericity (onlyFor3DROI) showed an AUC of 1.0; in the right posterior ocular fat, the texture CONVENTIONAL_TLG (mL) (onlyForPETorNM) had an AUC of 0.933, while the DISCRETIZED_TLG (mL) (onlyForPETorNM) displayed an AUC of 0.283. Other significant textures included SHAPE_Volume (mL) and SHAPE_Volume (vx) with AUCs of 1.0, and SHAPE_Sphericity (onlyFor3DROI) with an AUC of 0.1. Additionally, SHAPE_Surface (mm²) (onlyFor3DROI) achieved an AUC of 1.0, while GLRLM_RLNU and GLZLM_GLNU had AUCs of 0.933, and GLCM_Correlation had an AUC of 0.800. For the left retroocular fat, textures such as CONVENTIONAL_TLG (mL) and DISCRETIZED_TLG (mL) (both with AUCs of 0.933), as well as SHAPE_Volume (mL) and SHAPE_Volume (vx) (both with AUCs of 1.0), were noted. In the thyroid, CONVENTIONAL_SUVbwQ2 had an AUC of 0.933, and in the right lung, CONVENTIONAL_TLG (mL), DISCRETIZED_TLG (mL) (both with AUCs of 1.0), and SHAPE_Surface (mm²) (onlyFor3DROI) with an AUC of 0.833 were significant. In the left lung, notable textures such as SHAPE_Volume (mL) (AUC = 0.933), SHAPE_Volume (vx) (AUC = 0.933), and SHAPE_Surface (mm²) (onlyFor3DROI) (AUC = 0.967) were recorded, alongside SHAPE_Compacity (onlyFor3DROI) (AUC = 0.900), GLRLM_GLNU (AUC = 0.800), GLZLM_LZHGE (AUC = 0.867), and NGLDM_Busyness (AUC = 0.767). For the gastric sinus, CONVENTIONAL_SUVbwKurtosis and CONVENTIONAL_SUVbwExcessKurtosis (both with AUCs of 0.900), as well as SHAPE_Volume (mL) (AUC = 0.867), were significant. In the left kidney, GLRLM_RLNU had an AUC of 0.867, and for the L4 vertebral texture, GLZLM_GLNU showed an AUC of 0.867. There were no statistically significant differences in the textures of the thymus, aortic arch, left ventricle, liver, pancreas, right kidney, or spleen. The detailed results are shown in Table 3; Fig. 4.

Table 3.

ROC analysis results of PET differential texture between the young and elderly rhesus macaque groups

Organ	Radiomics Feature(RF)	AUC	Threshold
Brain	SHAPE_Sphericity[onlyFor3DROI]	1	0.718
R-Fat	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	0.933	1.931
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	0.933	7.165
	SHAPE_Volume(mL)	1	2
	SHAPE_Volume(vx)	1	250
	SHAPE_Sphericity[onlyFor3DROI])	0.1
	SHAPE_Surface(mm2)[onlyFor3DROI]	1	1636
	GLRLM_RLNU	0.933	126
	NGLDM_Coarseness	0.033
	GLZLM_GLNU	0.933	3.9
	GLCM_Correlation	0.8	0.455
L-Fat	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	0.933	1.942
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	0.933	7.153
	SHAPE_Volume(mL)	1	2.003
	SHAPE_Volume(vx)	1	251
Thyroid	CONVENTIONAL_SUVbwQ2	0.933	0.66
R-Lung	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	1	62
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	1	283
	SHAPE_Surface(mm2)[onlyFor3DROI]	0.833	26,022
L-Lung	SHAPE_Volume(mL)	0.933	128
	SHAPE_Volume(vx)	0.933	2934
	SHAPE_Surface(mm2)[onlyFor3DROI]	0.967	26,388
	SHAPE_Compacity[onlyFor3DROI]	0.9	5.818
	GLRLM_GLNU	0.8	530
	GLZLM_LZHGE	0.867	154,555
	NGLDM_Busyness	0.767	27.722
Stomach	CONVENTIONAL_SUVbwKurtosis	0.9	2.451
	CONVENTIONAL_SUVbwExcessKurtosis	0.9	-0.55
	SHAPE_Volume(mL)	0.867	1.5
L-Kidney	GLRLM_RLNU	0.867	157
L4	GLZLM_GLNU	0.867	3.451
Thymus
Aorta
Left Ventricular
Liver
Pancreas
R-Kidney
Spleen

Fig. 4.

ROC Curve of PET differential texture between the Young and Elderly Rhesus Macaque groups. A: Brain texture AUC = 1; B: Right retrobulbar fat texture, sequentially AUC = 0.933, AUC = 1, AUC = 1, AUC = 0.1, AUC = 1, AUC = 0.933, AUC = 0.033, AUC = 0.933, AUC = 0.800; C: Left retrobulbar fat texture, sequentially AUC = 0.933, AUC = 0.933, AUC = 1, AUC = 1; D: Thyroid texture AUC = 0.933; E: Right lung texture, sequentially AUC = 1, AUC = 1, AUC = 0.833; F: Left lung texture, sequentially AUC = 0.933, AUC = 0.933, AUC = 0.967, AUC = 0.900, AUC = 0.8, AUC = 0.867, AUC = 0.767; G: Gastric antrum texture, sequentially AUC = 0.900, AUC = 0.900, AUC = 0.867; H: Left kidney texture AUC = 0.867; L4 vertebra texture AUC = 0.867

PET texture evaluation of BMMSCs treatment efficacy in rhesus macaques

In the brain, the texture SHAPE_Sphericity [onlyFor3DROI] had a coefficient of 0.707 ± 0.010 in the young group and 0.735 ± 0.012 in the elderly group, with an evaluation threshold of 0.718. After 3 months of treatment in the elderly group, 4 animals showed a tendency towards the young group, and after 6 months of treatment, 4 animals also showed a tendency towards the young group. Significant differences were observed between the 3-month and 6-month post-treatment groups and the elderly group (P < 0.05), but no statistical difference was found between these groups and the young group.

Right retrobulbar fat texture analysis showed the following results: CONVENTIONAL_TLG (mL) [onlyForPETorNM] was 1.025 ± 0.613 in the young group and 2.408 ± 0.817 in the elderly group, with an evaluation threshold of 1.931. After 3 months of treatment, 1 elderly rhesus macaque tended to approach the young group, and after 6 months of treatment, 3 elderly macaques tended to approach the young group. The 3-month treatment group showed a significant difference compared to the young group (P < 0.05), while no significant difference was found between the 6-month treatment group and either the young or elderly groups. DISCRETIZED_TLG (mL) [onlyForPETorNM] was 3.217 (2.032, 6.155) in the young group and 9.213 (7.095, 10.911) in the elderly group, with an evaluation threshold of 7.165. After 3 months of treatment, 1 elderly rhesus macaque tended to approach the young group, and after 6 months of treatment, 3 elderly macaques tended to approach the young group. The 3-month treatment group showed a significant difference compared to the young group (P < 0.05), while no significant difference was found between the 6-month treatment group and either the young or elderly groups. SHAPE_Volume (mL) was 1.325 (0.97, 1.669) in the young group and 2.778 (2.387, 3.245) in the elderly group, with an evaluation threshold of 2. After 3 months of treatment, no significant change was observed, and after 6 months of treatment, 3 elderly macaques tended to approach the young group. The 3-month post-treatment group showed a significant difference compared to the young group (P < 0.05), while no significant difference was found between the 6-month treatment group and the young group, but a significant difference was found compared to the elderly group (P < 0.05). SHAPE_Volume (vx) was 166 (121, 209) in the young group and 348 (299, 407) in the elderly group, with an evaluation threshold of 250. After 3 months of treatment, no significant change was observed, and after 6 months of treatment, 3 elderly macaques tended to approach the young group. The 3-month post-treatment group showed a significant difference compared to the young group (P < 0.05), while the 6-month treatment group showed a significant difference compared to the elderly group (P < 0.05). SHAPE_Surface (mm²) [onlyFor3DROI] was 1194 ± 317 in the young group and 2346 ± 468 in the elderly group, with an evaluation threshold of 1636. After 3 months of treatment, no significant change was observed, and after 6 months of treatment, 3 elderly macaques tended to approach the young group. The 3-month post-treatment group showed a significant difference compared to the young group (P < 0.05), while no significant difference was found between the 6-month treatment group and either the young or elderly groups. GLRLM_RLNU was 85 (67, 102) in the young group and 170 (128, 200) in the elderly group, with an evaluation threshold of 126. No significant change was observed in the 3-month and 6-month treatment groups. GLZLM_GLNU was 3.364 (2.929, 4) in the young group and 4.776 (4.498, 5.058) in the elderly group, with an evaluation threshold of 3.9. After 3 months of treatment, 3 elderly macaques tended to approach the young group, and after 6 months of treatment, 4 elderly macaques tended to approach the young group. The 3-month treatment group showed a significant difference compared to the elderly group (P < 0.05), while no significant difference was found between the 6-month treatment group and the elderly group. GLCM_Correlation was 0.431 ± 0.143 in the young group and 0.597 ± 0.141 in the elderly group, with an evaluation threshold of 0.455. Both the 3-month and 6-month post-treatment groups showed a significant difference compared to the young group (P < 0.05), while no significant difference was found between the post-treatment groups and the elderly group.

Left retrobulbar fat texture analysis showed the following results: CONVENTIONAL_TLG (mL) [onlyForPETorNM] was 1.066 (0.607, 1.643) in the young group and 2.462 (1.927, 2.927) in the elderly group, with an evaluation threshold of 1.942. After 3 months of treatment, no elderly rhesus macaques tended to approach the young group, but after 6 months of treatment, 3 elderly macaques tended to approach the young group. The 3-month and 6-month post-treatment groups showed a significant difference compared to the young group (P < 0.05), with no significant difference compared to the elderly group. DISCRETIZED_TLG (mL) [onlyForPETorNM] was 4.064 (2.431, 6.083) in the young group and 9.349 (7.303, 11.129) in the elderly group, with an evaluation threshold of 7.153. After 3 months of treatment, no elderly rhesus macaques tended to approach the young group, but after 6 months of treatment, 3 elderly macaques tended to approach the young group. The 3-month and 6-month post-treatment groups showed a significant difference compared to the young group (P < 0.05), with no significant difference compared to the elderly group. SHAPE_Volume (mL) was 1.289 ± 0.322 in the young group and 3.048 ± 0.832 in the elderly group, with an evaluation threshold of 2.003. After 3 months of treatment, 1 elderly rhesus macaque tended to approach the young group, and after 6 months of treatment, 4 elderly macaques tended to approach the young group. The 3-month and 6-month post-treatment groups showed a significant difference compared to the young group (P < 0.05), with no significant difference compared to the elderly group. SHAPE_Volume (vx) was 161 ± 40 in the young group and 382 ± 104 in the elderly group, with an evaluation threshold of 251. After 3 months of treatment, 1 elderly rhesus macaque tended to approach the young group, and after 6 months of treatment, 4 elderly macaques tended to approach the young group. The 3-month and 6-month post-treatment groups showed a significant difference compared to the young group (P < 0.05), with no significant difference compared to the elderly group.

In the thyroid, the texture CONVENTIONAL_SUVbwQ2 was 0.588 (0.535, 0.673) in the young group and 0.901 (0.683, 0.995) in the elderly group, with an evaluation threshold of 0.66. After 3 months of treatment, no elderly rhesus macaques tended to approach the young group, and after 6 months of treatment, no elderly macaques tended to approach the young group. The 3-month and 6-month post-treatment groups showed a significant difference compared to the young group (P < 0.05), with no significant difference compared to the elderly group.

The analysis of right lung texture revealed the following results: CONVENTIONAL_TLG (mL) [onlyForPETorNM] was 50 (36, 59) in the young group and 69 (66, 91) in the elderly group, with a threshold of 62. Both the 3-month and 6-month post-treatment groups exhibited significant differences compared to the young group (P < 0.05), but no significant difference was found when compared to the elderly group. DISCRETIZED_TLG (mL) [onlyForPETorNM] was 229 (155, 258) in the young group and 294 (289, 368) in the elderly group, with a threshold of 283. The 3-month post-treatment group showed a significant difference when compared to the young group, while the 6-month post-treatment group showed significant differences when compared to both the young and elderly groups (P < 0.05). No significant difference was observed between the 3-month post-treatment and elderly groups. The SHAPE_Surface (mm²) [onlyFor3DROI] was 21,630 ± 6,141 in the young group and 27,464 ± 2,944 in the elderly group, with a threshold of 26,022. The 3-month post-treatment group exhibited a significant difference when compared to both the young and elderly groups (P < 0.05), while the 6-month post-treatment group showed a significant difference compared to the young group, but no significant difference compared to the elderly group.

The analysis of left lung texture revealed the following results: SHAPE_Volume (mL) was 109 (106, 139) in the young group and 172 (155, 186) in the elderly group, with a threshold of 128. Both the 3-month and 6-month post-treatment groups showed significant differences compared to the young group (P < 0.05), but no significant difference compared to the elderly group. SHAPE_Volume (vx) was 2,501 (2,443, 3,196) in the young group and 3,966 (3,564, 4,283) in the elderly group, with a threshold of 2,934. Both the 3-month and 6-month post-treatment groups showed significant differences compared to the young group (P < 0.05), with no significant difference compared to the elderly group. SHAPE_Surface (mm²) [onlyFor3DROI] was 21,456 ± 2,862 in the young group and 27,197 ± 2,295 in the elderly group, with a threshold of 26,388. After 3 months, two elderly macaques and after 6 months, one elderly macaque approached the values seen in the young group. Both the 3-month and 6-month post-treatment groups showed significant differences compared to the young group (P < 0.05), with no significant difference compared to the elderly group. SHAPE_Compacity [onlyFor3DROI] was 5.55 ± 0.532 in the young group and 6.305 ± 0.432 in the elderly group, with a threshold of 5.818. The 3-month post-treatment group showed a significant difference compared to the young group, and the 6-month post-treatment group showed significant differences compared to both the young and elderly groups (P < 0.05). No significant difference was observed between the 3-month post-treatment group and the elderly group. GLRLM_GLNU was 436 (390, 481) in the young group and 580 (484, 622) in the elderly group, with a threshold of 530. The 3-month post-treatment group showed significant differences compared to the young group (P < 0.05), while the 6-month post-treatment group showed no significant difference compared to either the young or elderly groups. GLZLM_LZHGE was 85,839 (51,603, 211,465) in the young group and 345,988 (176,262, 468,263) in the elderly group, with a threshold of 154,555. After 3 months, five elderly macaques, and after 6 months, two elderly macaques showed values approaching those of the young group. The 3-month post-treatment group showed a significant difference compared to the elderly group (P < 0.05), but no significant difference compared to the young group. The 6-month post-treatment group showed no significant difference compared to either the young or elderly groups. NGLDM_Busyness was 14.147 ± 7.690 in the young group and 24.213 ± 10.298 in the elderly group, with a threshold of 27.722. After 3 months, six elderly macaques, and after 6 months, five elderly macaques tended to approach the values seen in the young group. Both the 3-month and 6-month post-treatment groups showed significant differences compared to the elderly group (P < 0.05), but no significant difference compared to the young group.

The gastric antrum texture analysis revealed the following results: CONVENTIONAL_SUVbwKurtosis was 2.224 ± 0.636 in the young group and 3.263 ± 0.769 in the elderly group, with a threshold of 2.451. After 3 months of treatment, three elderly macaques, and after 6 months, five elderly macaques showed values approaching those of the young group. The 6-month post-treatment group demonstrated a significant difference compared to the elderly group (P < 0.05), but no significant difference compared to the young group. The 3-month post-treatment group showed no significant difference compared to either the young or elderly groups. CONVENTIONAL_SUVbwExcessKurtosis was − 0.776 ± 0.636 in the young group and 0.263 ± 0.769 in the elderly group, with a threshold of -0.55. After 3 months of treatment, three elderly macaques, and after 6 months, five elderly macaques showed values approaching those of the young group. The 6-month post-treatment group demonstrated a significant difference compared to the elderly group (P < 0.05), but no significant difference compared to the young group. The 3-month post-treatment group showed no significant difference compared to either the young or elderly groups. SHAPE_Volume (mL) was 0.391 (0.283, 2.021) in the young group and 2.347 (2.032, 3.673) in the elderly group, with a threshold of 1.5. After 3 months of treatment, four elderly macaques, and after 6 months, three elderly macaques showed values approaching those of the young group. Both the 3-month and 6-month post-treatment groups showed significant differences compared to the elderly group (P < 0.05), with no significant difference compared to the young group.

The texture of the left kidney, measured by GLRLM_RLNU, was 111 (65, 167) in the young group and 194 (157, 362) in the elderly group, with an evaluation threshold of 157. Both the 3-month and 6-month post-treatment groups showed a significant difference compared to the young group (P < 0.05), but no significant difference compared to the elderly group.

In the L4 vertebra, the texture GLZLM_GLNU was 2.137 ± 0.595 in the young group and 4.056 ± 1.135 in the elderly group, with an evaluation threshold of 3.452. After 3 months of treatment, 3 elderly monkeys showed a tendency to approach the young group, and after 6 months of treatment, 5 elderly monkeys showed a similar trend. The 6-month treatment group demonstrated a significant difference compared to the elderly group (P < 0.05), but no significant difference compared to the young group. No statistical differences were observed between the 3-month treatment group and either the young or elderly groups, as shown in Table 4.

Table 4.

Efficacy evaluation of BMMSCs treatment in aged rhesus monkeys

Organ	Radiomics Feature(RF)	Young	Elderly	Threshold	After-3 M	After-6 M
Brain	SHAPE_Sphericity[onlyFor3DROI]	0.707 ± 0.010	0.735 ± 0.012	0.718	4/6(66.7%)	4/6(66.7%)
R-Fat	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	1.025 ± 0.613	2.408 ± 0.817	1.931	1/6(16.7%)	3/6(50%)
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	3.217(2.032,6.155)	9.213(7.095,10.911)	7.165	1/6(16.7%)	3/6(50%)
	SHAPE_Volume(mL)	1.325(0.97,1.669)	2.778(2.387,3.245)	2	0/6(0%)	3/6(50%)
	SHAPE_Volume(vx)	166(121,209)	348(299,407)	250	0/6(0%)	3/6(50%)
	SHAPE_Sphericity[onlyFor3DROI])	0.46(0.458,0.549)	0.415(0.402,0.442)
	SHAPE_Surface(mm2)[onlyFor3DROI]	1194 ± 317	2346 ± 468	1636	0/6(0%)	3/6(50%)
	GLRLM_RLNU	85(67,102)	170(128,200)	126	0/6(0%)	4/6(66.7%)
	NGLDM_Coarseness	0.053 ± 0.014	0.028 ± 0.007
	GLZLM_GLNU	3.364(2.929,4)	4.776(4.498,5.058)	3.9	3/6(50%)	4/6(66.7%)
	GLCM_Correlation	0.431 ± 0.143	0.596 ± 0.141	0.455	0/6(0%)	0/6(0%)
L-Fat	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	1.066(0.607,1.643)	2.462(1.927,2.927)	1.942	0/6(0%)	3/6(50%)
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	4.064(2.431,6.083)	9.349(7.303,11.129)	7.153	0/6(0%)	3/6(50%)
	SHAPE_Volume(mL)	1.289 ± 0.322	3.048 ± 0.832	2.003	1/6(16.7%)	4/6(66.7%)
	SHAPE_Volume(vx)	161 ± 40	382 ± 104	251	1/6(16.7%)	4/6(66.7%)
Thyroid	CONVENTIONAL_SUVbwQ2	0.588(0.0.535,0.673)	0.901(0.683,0.995)	0.66	0/6(0%)	0/6(0%)
R-Lung	CONVENTIONAL_TLG(mL)[onlyForPETorNM]	50(36,59)	69(66,91)	62	0/6(0%)	0/6(0%)
	DISCRETIZED_TLG(mL)[onlyForPETorNM]	229(155,258)	294(289,368)	283	0/6(0%)	0/6(0%)
	SHAPE_Surface(mm2)[onlyFor3DROI]	21,630 ± 6141	27,464 ± 2944	26,022	0/6(0%)	0/6(0%)
L-Lung	SHAPE_Volume(mL)	109(106,139)	172(155,186)	128	0/6(0%)	0/6(0%)
	SHAPE_Volume(vx)	2501(2443,3196)	3966(3564,4283)	2934	0/6(0%)	0/6(0%)
	SHAPE_Surface(mm2)[onlyFor3DROI]	21,456 ± 2862	27,197 ± 2295	26,388	2/6(33.3%)	1/6(16.7%)
	SHAPE_Compacity[onlyFor3DROI]	5.55 ± 0.532	6.305 ± 0.432	5.818	0/6(0%)	0/6(0%)
	GLRLM_GLNU	436(390,481)	580(484,622)	530	0/6(0%)	1/6(16.7%)
	GLZLM_LZHGE	85,839(51603,211465)	345,988(176262,468263)	154,555	5/6(83.3%)	2/6(33.3%)
	NGLDM_Busyness	14.147 ± 7.690	24.213 ± 10.298	27.722	6/6(100%)	6/6(100%)
Stomach	CONVENTIONAL_SUVbwKurtosis	2.224 ± 0.636	3.263 ± 0.769	2.451	3/6(50%)	5/6(83.3%)
	CONVENTIONAL_SUVbwExcessKurtosis	-0.776 ± 0.636	0.263 ± 0.769	-0.55	3/6(50%)	5/6(83.3%)
	SHAPE_Volume(mL)	0.391(0.283,2.021)	2.347(2.032,3.673)	1.5	4/6(66.7%)	3/6(50%)
Kidney-L	GLRLM_RLNU	111(65,167)	194(157,362)	157	0/6(0%)	0/6(0%)
L4	GLZLM_GLNU	2.137 ± 0.595	4.056 ± 1.135	3.452	5/6(83.3%)	5/6(83.3%)
Thymus
Aorta
Left Ventricular
Liver
Pancreas
R-Kidney
Spleen

HE staining results showed that after 6 months of BMMSC treatment, the histopathological structures of various organ tissues were significantly improved. In the juvenile and young groups, alveolar structures were clear and intact, with thin alveolar walls; liver lobules were well-defined, and hepatocyte cords were neatly arranged; glomeruli were full, and renal tubule structures were clear; gastric mucosa was well-layered, and glands were regular; the thymus exhibited clear boundaries between the cortex and medulla, with high lymphocyte density, all demonstrating typical young organ morphology. In contrast, in the aged group, the alveolar septa were widened and fused with mild inflammatory infiltration; hepatocyte cords were slightly disordered, blood sinusoids were dilated with lipofuscin deposition, and the portal area exhibited increased collagen and mononuclear cell infiltration; glomeruli were atrophied or sclerotic, afferent arterioles were thickened, renal tubules were dilated and atrophied with interstitial fibrosis; gastric mucosa was thinned, glands were reduced or atrophied with inflammatory cell infiltration; the thymus showed significant atrophy, with blurred cortex and medulla boundaries, replaced by adipose tissue and lymphocyte reduction, all displaying features of aging or functional degeneration. After 6 months of stem cell treatment, the aged rhesus monkeys showed more orderly alveolar structures, restored hepatic cell arrangement, glomeruli and renal tubule morphology resembling those of the young group, well-preserved gastric mucosa and glandular structures, clear thymus cortex and medulla boundaries, increased lymphocyte density, and reduced fat infiltration, demonstrating significant improvements in the structure of multiple organ tissues. These results are shown in Fig. 5.

Fig. 5.

Histological comparison of major organs at different age stages and after BMMSC treatment. The figure shows hematoxylin and eosin (H&E) stained tissue sections of major organs in different groups. A-E represent the lung, liver, kidney, stomach, and thymus, respectively. Each column, from left to right, displays a morphological comparison of tissues from juvenile, young, elderly, and elderly individuals after 6 months of BMMSC treatment

Discussion

In the brain, significant differences in spherical shape features based on 3D ROIs were observed between the young and aged groups. After BMMSC treatment, the spherical shape feature texture in the aged rhesus monkeys gradually approached that of the young group after 3 and 6 months, suggesting an improvement in brain function. Specifically, the structure and function of brain regions such as the frontal cortex and hippocampus, which are involved in cognitive and memory functions, showed significant changes following treatment [18, 19]. The mechanism may involve both the direct cellular integration and paracrine effects of BMMSCs. El-Akabawy et al. demonstrated that transplanted BMMSCs not only differentiate into neural cells but also secrete paracrine factors, such as brain-derived neurotrophic factor (BDNF) and nerve growth factor (NGF), which promote synaptic remodeling and the restoration of the cholinergic system, thereby improving motor activity and cognitive function in senescent rats [20].

In the right retroocular fat, texture features based on TLG, shape-volume, precision, and surface area exhibited significant differences between the youth and elderly groups. After BMMSCs treatment, the run-length homogeneity of the length matrix in the grayscale region showed a tendency to converge towards the youth group in the elderly group at 3 months post-treatment, while 5 texture features gradually converged towards the youth group at 6 months post-treatment. In the left retroocular fat, TLG and shape-volume features also showed significant differences, with 4 texture features in the elderly group gradually converging towards those of the younger group at 6 months post-treatment. The redistribution of adipose tissue during aging, the decline in the function of adipose progenitor and stem cells, and the aging of adipocytes themselves may be the primary factors contributing to the differences in adipose tissue texture characteristics between the elderly and youth groups. After 6 months of BMMSCs treatment, the function of adipose progenitor and stem cells in the retroocular adipose tissue of aged rhesus monkeys was replenished, the redistribution trend of adipose tissue was reversed, and the efficacy of the 6-month treatment group was significantly better than that of the 3-month treatment group [21, 22]. On the one hand, the transplanted MSCs can home to the target tissue and differentiate into adipocytes, thereby contributing to tissue remodeling and repairing functional defects [23]. On the other hand, stem cells can enhance the local microenvironment and maintain the dynamic balance of fat metabolism by secreting paracrine factors such as vascular endothelial growth factor (VEGF), fibroblast growth factor (FGF), and other regulatory molecules [24]. This finding further supports the potential use of BMMSCs in ameliorating age-related changes in adipose tissue.

In the thyroid gland, there was a significant difference in conventional SUVbwQ2 texture between the young and elderly groups. After BMMSCs treatment, the texture of the elderly group did not significantly converge to that of the young group at both 3 and 6 months. The study suggests that under certain conditions, stem cells can differentiate into normal thyroid cells and restore thyroid function [25], This provides new possibilities for the treatment of thyroid-related diseases. Additionally, by transplanting human amniotic MSCs, thyroid cell apoptosis can be inhibited in aged mice, and their immune function can be enhanced, thereby improving subclinical hypothyroidism. While FDG-PET standardized uptake values can be used to predict thyroid disorders, their specificity remains low [26, 27]. In this study, the texture features of the thyroid based on SUV did not show significant changes in aged rhesus monkeys after BMMSCs treatment.

In the right lung, significant differences in conventional and discretized TLG, as well as shape surface area texture features, were observed between the young and elderly groups. After BMMSCs treatment, the texture features in the elderly group did not significantly converge to those of the young group at both 3 and 6 months. In the left lung, texture features such as shape volume, surface area, and compactness also showed significant differences between the young and elderly groups. After treatment, some texture features in the elderly group tended to approach those of the young group at 3 months, while at 6 months, the frequency texture of the neighborhood gray-level difference matrix converged to the young group. This improvement may be closely related to the paracrine effects of BMMSCs. The secretion of transforming growth factor-β (TGF-β) and interleukin-10 (IL-10) by BMMSCs, as well as the regulation of immune response through the increase in immunosuppressive T regulatory cells (Tregs), together help alleviate inflammation and promote lung tissue repair. Furthermore, the Wnt/β-catenin signaling pathway plays an important role in the development and repair of various tissues. Stem cells may promote the differentiation of lung-resident mesenchymal stem cells (LR-MSCs) into epithelial cells through the activation of the Wnt/β-catenin signaling pathway [28]. This process may play a crucial role in the recovery after lung injury, particularly in the treatment of diseases such as pulmonary fibrosis, providing a potential stem cell therapy strategy. Our previous research showed that intravenous infusion of BMMSCs led to a reduction in the degree of pulmonary fibrosis in elderly rhesus monkeys, an increase in the density of capillaries around the alveoli (p < 0.05), and a significant increase in the number of type II alveolar epithelial cells (p < 0.05). However, no significant changes were observed in the SUVmax-based evaluation [29]. The high-intensity regions in the gray-level run-length matrix and the frequency in the neighborhood gray-level difference matrix may serve as effective indicators for pre-assessing pulmonary changes following BMMSCs treatment.

In the gastric sinus, significant differences in kurtosis, hyperkurtosis, and shape-volume textures of conventional SUVs were observed between the young and aged groups. After BMMSCs treatment, these textures in the aged rhesus monkeys tended to converge toward those of the young group at both 3 and 6 months, suggesting an improvement in gastric sinus function. This improvement may be attributed to BMMSCs promoting the proliferation of gastrointestinal epithelial cells via paracrine secretion of epidermal growth factor (EGF), as well as directly migrating to the site of mucosal injury and differentiating into functional cells to restore the integrity of the gastric sinus mucosa [30, 31]. Relevant studies have demonstrated that BMMSCs can enhance the repair of gastrointestinal epithelial cell damage [32, 33], This provides strong support for the clinical application of BMMSCs in treating gastrointestinal diseases.

In the left kidney, the run length inhomogeneity texture of the grayscale length run matrix was significantly different between the youth and aged groups. However, after BMMSCs treatment, the texture changes in the aged rhesus monkeys did not significantly converge to those of the young group after 3 and 6 months. Although BMMSCs have been shown to significantly alleviate acute kidney injury and chronic renal failure, their mechanisms may involve differentiation into renal cells, contributing to renal repair and regeneration. Additionally, BMMSCs may reduce tissue fibrosis by decreasing TGF-β1 expression and attenuate inflammation by inhibiting HDAC2 expression [34, 35]. However, the sensitivity of PET texture analysis to changes in renal function remains limited, likely due to the complexity of renal tissue, spatial resolution constraints, and the dynamic nature of functional recovery, which prevents significant changes in texture.

In L4 cones, the run-length uniformity texture of the length matrix in the gray-scale region showed significant differences between the youth and aged groups. After BMMSCs treatment, the texture of the rhesus monkeys in the aged group tended to be higher than that of the young group at both 3 and 6 months. This effect on bone tissue regeneration may arise from the direct osteogenic differentiation capacity of BMMSCs and the secretion of factors such as bone morphogenetic proteins (BMPs) and alkaline phosphatase (ALP), which promote bone differentiation and work synergistically to enhance bone deposition and microstructural reconstruction. Over the past decade, research on the efficacy of bone marrow-derived mesenchymal stem/stromal cells (BMSCs) in bone tissue engineering and regenerative medicine has steadily increased. Further studies have shown that BMMSCs not only participate in bone regeneration through direct differentiation but also release various cytokines and growth factors, such as vascular endothelial growth factor (VEGF) and transforming growth factor-β (TGF-β), via paracrine mechanisms. These factors promote local angiogenesis, inhibit inflammatory responses, and accelerate the repair process of damaged areas [36–39]. These mechanisms work synergistically to enhance the potential of BMMSCs in bone tissue regeneration, making them a promising strategy for treating bone injuries and diseases.

Our study demonstrated that stem cell treatment in aged rhesus monkeys led to a convergence of tissue transcriptome expression towards the thymic profile of young rhesus monkeys, effectively reversing the expression of aging-related proteins [9]. Iyeon Kim’s study found that stem cell transplantation improved clinical parameters associated with cardiac disease in patients with heart disease [40]. Chenxia Hu pointed out that mesenchymal stem cells were able to migrate to damaged tissues, induce hepatogenic differentiation, inhibit the release of inflammatory factors, and enhance hepatocyte proliferation [41]. Naomi D’Souza demonstrated that stem cells can differentiate into insulin-secreting cells, thereby ameliorating hyperglycemia and addressing other pancreatic disorders in diabetic rats [42]. Yu-Liang Sun demonstrated that MSC transplantation alleviated splenic injuries in mice with type 1 diabetes (T1D) [43]. Jun Chen demonstrated that BMSC transplantation promotes renal tubular regeneration in mice with acute renal ischemia [44]. Several studies have demonstrated that stem cell transplantation promotes cell regeneration and functional recovery in organs such as the thymus, blood vessels, myocardium, liver, pancreas, spleen, and kidney. However, the efficacy of transplantation is not yet clearly observable in our PET images, possibly due to the gradual process of cellular integration post-transplantation or because cellular regeneration has not yet reached a level detectable by the current imaging resolution.

In summary, aging signs in the brain, left lung, gastric antrum, and L4 vertebra of elderly rhesus monkeys treated with BMMSCs were significantly alleviated at both 3 and 6 months post-treatment, suggesting the potential of BMMSCs for multi-organ anti-aging effects. Notably, a redistribution of retro-orbital fat was observed at 6 months after treatment, indicating that BMMSCs may also have profound effects on adipose tissue metabolism. Histological evaluation using H&E staining in the lung, liver, kidney, stomach, and thymus revealed marked structural improvements in the lung, kidney, and stomach. Although partial structural restoration was observed in the liver and thymus, PET imaging did not reveal significant textural changes in these organs, implying a possible temporal delay or sensitivity difference between histological recovery and functional imaging. PET texture analysis, as a quantifiable imaging technique reflecting both tissue structure and function, demonstrated high sensitivity and reproducibility in detecting BMMSC-induced recovery in the brain, lungs, and adipose tissue. Moreover, it not only captures local metabolic and structural changes but can also be integrated with genomics and proteomics data to uncover underlying molecular mechanisms, paving the way for a more precise therapeutic evaluation system [45]. Although this study is still at the stage of animal experimentation, the results provide a solid theoretical foundation and data support for the potential clinical application of BMMSCs in the treatment of age-related diseases. The significant improvement in multiple aging-related indicators observed in the rhesus monkey model suggests that BMMSCs may serve as a promising therapeutic strategy for intervening in human aging. However, translating these findings effectively into clinical practice remains challenging and requires targeted efforts in future research. First, it is essential to conduct systematic dose–response studies to determine the optimal therapeutic dose and administration route of BMMSCs in humans [46–48]. Secondly, although no severe adverse effects were observed in the non-human primate model used in this study, it remains critical to closely monitor potential tumorigenic risks and long-term safety in human clinical trials [46, 49]. Future clinical studies should prioritize the evaluation of the efficacy and safety of BMMSCs in specific age-related diseases. BMMSCs hold broad application prospects in various aging-associated conditions, such as neurodegenerative diseases and pulmonary dysfunction [50], PET texture analysis offers a reliable and precise imaging assessment tool for early efficacy identification, dynamic monitoring, patient stratification, and individualized dose optimization, thereby facilitating the precise implementation and clinical translation of BMMSC-based therapies. The integration of multimodal imaging with histological evaluation provides a more comprehensive understanding of both the short-term and long-term effects of BMMSC treatment, laying a solid and feasible technical foundation for its application in clinical translation and precision medicine. Future studies should focus on addressing the key challenges in clinical translation, including the optimization of cell preparation and quality control processes, the establishment of standardized therapeutic protocols, and the implementation of multicenter, large-sample clinical trials, ultimately advancing the broad application of BMMSCs in the treatment of aging and related diseases.

This study has several limitations. First, the use of rhesus monkeys with a small sample size limits the ability to train statistics-based machine learning algorithms and develop predictive models. Second, the analysis relied solely on imaging and histology data. Future studies could enhance the comprehensiveness and accuracy of the findings by incorporating additional data, such as clinical rating scales and serological markers. Finally, this study did not analyze image texture in relation to the transcriptome expression of whole-body organs and tissues, nor did it delve into imaging genomics. Future research should explore the connection between imaging and genomic data to offer a more comprehensive understanding of the mechanisms underlying stem cell therapy.

Conclusion

¹⁸F-FDG-PET textural features effectively distinguished differences between young and aged rhesus monkeys in organs such as the brain, bilateral retro-ocular fat, thyroid, both lungs, gastric sinus, left kidney, and L4 vertebrae.

¹⁸F -FDG-PET textural features serve as a powerful tool for assessing BMMSC efficacy. At 3 and 6 months post-treatment, aged rhesus monkeys showed significant improvement in textural features of the brain, bilateral retro-ocular fat, left lung, gastric sinus, and L4 vertebrae, suggesting a reversal of the aging process.

Multiple organs (e.g., the brain, left lung, gastric sinus, and L4 vertebrae) in aged rhesus monkeys demonstrated a delayed aging process at 3 and 6 months after BMMSC treatment, with redistribution of retro-ocular fat beginning at 6 months.

HE staining revealed significant improvements in the histopathological structures of the lungs, liver, kidneys, stomach, and thymus after 6 months of BMMSC treatment, supporting the imaging findings at the histological level.

¹⁸F -FDG-PET textural characterization provides an effective method for in vivo tissue and organ function assessment.

Author contributions

Gaohong Zhu and Xinghua Pan designed the study; Shici Yang and Rui He performed the image acquisitions; Zhiyu Yang and Yuanyuan Li performed the image analysis; Yaohui Zhang and Jie Li interpreted the results and wrote the initial manuscript; Longbao Lü and Xiangqing Zhu provided technical and theoretical support for this study to address the challenges encountered.all authors contributed to and approved the final manuscript.

Funding

This study was supported by the National Natural Science Foundation of China (Project No. 82260354, “11C-labeled GSK-3β in Neuroimaging and Molecular Mechanism of Cognitive Impairment Secondary to Chronic Lymphocytic Thyroiditis,” 2023/01–2026/12), and the Yunnan Provincial High-level Talent Program – “Xingdian Talent Support Program” (Distinguished Physician Category, Project No. RLMY20200010, 2021/01–2025/12).

Data availability

All research data are available from the corresponding author upon request.

Declarations

Ethical approval

All experimental animals were housed at the Kunming Institute of Zoology, Chinese Academy of Medical Sciences [SYXK(Yunnan)K2017-0008]. The housing and experimental procedures strictly followed the guidelines of the Public Health Service Policy on Humane Care and Use of Laboratory Animals (USA). The animals were kept in an environment with a controlled light-dark cycle (daytime: 8:00–20:00, nighttime: 20:00–8:00), at a temperature of 25–27 °C and humidity of 55–70%. This study was approved by the Ethics Committee of the 920th Hospital of the Joint Logistics Support Force of the People’s Liberation Army of China (Approval No. Lunshen2019-003(Ke)-01).

The title of the approved project is “Stem cell technology and clinical translational research,” and the date of approval is June 27, 2018.

The registration number of the certificate of accreditation for the experimental animal facility is CNAS LA0001.

Use of artificial intelligence

The authors declare that they have not used AI-generated work in this manuscript.

Conflict of interest

The authors declare that they have no conflict of interest with regard to this study.