Testicular torsion diagnosis and injury assessment using photoacoustic oxygenation imaging

Abstract

Testicular torsion (TT) is a medical emergency that requires immediate diagnostic evaluation. Photoacoustic imaging (PAI) has the potential to provide spatially resolved oxygen saturation (sO₂), which can serve as a valuable marker in TT diagnosis. We investigated the potential of PAI as an alternative method for TT diagnosis and testicular injury assessment. We measured sO₂ levels in different degrees of TT models using PAI at various time points. Based on histopathological results, we found that the averaged sO₂ per pixel ($\stackrel{\circledR }{s{O}_2}$) and reduction of $\stackrel{\circledR }{s{O}_2}$ (rsO₂) in twisted testicles had significant correlations with hypoxic conditions. Both $\stackrel{\circledR }{s{O}_2}$ and rsO₂ exhibited excellent diagnostic abilities in detecting TT and identifying ischemia/hypoxia injury following TT. Furthermore, PAI-measured sO₂ demonstrated favorable diagnostic capabilities in discriminating if the testicle had suffered irreversible injury. In summary, PAI presents a potentially promising novel approach in evaluating TT and warrants further clinical investigation.

1. Introduction

Testicular torsion (TT) is a common urological emergency in children, where the blood flow to the testicles is obstructed due to the twisting of the spermatic cord. It is estimated that TT occurs in approximately 1 in 4000 males under the age of 25 annually, accounting for 25–35% of acute scrotum cases in children [1], [2], [3].

In clinical practice, the decisions to perform detorsion or orchidectomy depends on the degree of ischemia and necrosis. Studies have shown that salvaging the testicle is unlikely in most cases when torsion lasts for more than 6 h, making orchiectomy the recommended option [4]. However, in addition to the duration of torsion, the degree of rotation is another factor associated with testicular ischemia and clinical outcomes [3], [5], [6], [7], [8]. Therefore, timely detection of TT and precise assessment of testicular viability within the recommended 24-hour treatment window are crucial in guiding prompt clinical intervention and selecting appropriate treatment modalities.

Ultrasonography (US) is currently the preferred imaging modality for evaluating TT in clinical emergencies. The presence of the whirlpool sign and absence of blood flow on Colour Doppler flow imaging (CDFI) or grey-scale US are highly indicative of TT, as extensively documented [9], [10]. However, conventional US has inherent limitations, heavily relying on the expertise of sonographers and sensitivity of devices, especially in the pediatric population with small testes characterized by low vascular density and slow blood flow velocity [11], [12]. Additionally, in cases of partial torsion, the preservation of blood perfusion in the testes may lead to misdiagnosis and elevate the likelihood of false negatives [13], [14], [15], [16]. Contrast-enhanced ultrasound (CEUS) has emerged as the gold standard for diagnosing TT in recent years [17], [18]. However, the high cost of US contrast agents, requirement of intravenous injection, and safety concerns associated with their use in pediatric patients have restricted the application of CEUS in emergency settings [19], [20].

Photoacoustic imaging (PAI) is a rapidly developing biomedical imaging technique that is based on the photoacoustic (PA) effect. This non-invasive hybrid technique combines the high contrast of optical imaging with the high spatial resolution of ultrasound imaging to reconstruct structural or functional images of tissues from captured ultrasound signals, known as photoacoustic signals (PAS). These signals are evoked from endogenous chromophores, such as hemoglobin, melanin, and lipids, which are excited by pulsed laser light [21], [22], [23]. Hemoglobin is the predominant substance responsible for strong optical absorption in blood vessels within the range of light wavelengths from 690 to 850 nm. By analyzing the distinct optical absorption peaks of oxygenated hemoglobin (HbO₂) and deoxygenated hemoglobin (HbR), for example, at 840 nm and 760 nm, respectively [24], [25], PAI demonstrates the potential to accurately quantify levels of oxygenation and provide spatially resolved measurements of oxygen saturation (sO₂). This technique is label-free and non-invasive because the laser fluence falls within the safety range [26]. Studies have developed PAI measurements of sO₂ for evaluating tumor microenvironments, characterizing renal function, assessing myocardial ischemia, and examining ischemia-reperfusion in fingers [27], [28], [29], [30]. In these studies, HbO₂ and HbR have proven to be valuable markers for differentiating between normal and pathological tissues [31], [32].

TT can result in ischemia, hypoxia, and ultimately necrosis. The hypoxia stimulation may lead to an increase in the expression of hypoxia-inducible factor 1-alpha (HIF-1α) and a decrease in tissue sO₂ levels [33], [34]. In theory, sO₂ measurements obtained through PAI could potentially serve as a diagnostic tool for TT and detect changes in hemoglobin redox states during clinical emergencies. However, there have been no studies to date that have investigated the usefulness of PAI in evaluating TT.

The main aim of this study is to investigate the feasibility of using PAI to detect TT early and estimate the degree of testicular injury following TT, based on histopathological changes.

2. Methods

2.1. Animals

A total of 50 male New Zealand rabbits weighing between 2.2 and 3.0 kg were provided and housed under standard living conditions for a 2-week acclimatization period. All animal experiments were conducted in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Animal Care and Use Committee of West China Hospital (License Number: 20220519012).

2.2. TT modeling

Surgery was performed on rabbits after they were anesthetized with Zoletil at a dosage of 5 mg/kg intramuscularly, followed by intraperitoneal administration of 1% pentobarbital sodium at a dosage of 3 ml/kg. Prior to the surgery, the inguinal and scrotal regions were shaved and sterilized under aseptic conditions. Incisions were made at a distance of 1–2 cm above the scrotum, while maintaining sterile conditions. The spermatic cord was rotated in a counter-clockwise direction and subsequently affixed to the muscular wall at the same level. The animals were subjected to torsion at 0° (sham operation, group S; n = 5), 180° (group A; n = 15), 360° (group B; n = 15), and 720° (group C; n = 15) in a randomized manner. For the torsion groups (group A, B, C), rabbits were further divided into three subgroups (n = 5 for each subgroup) based on the time of sacrifice after the operation (2 h, 4 h, 6 h). Postoperative analgesia was administered intramuscularly with 0.2 ml/kg of meloxicam.

2.3. Colour Doppler ultrasound evaluation

Ultrasound examinations were performed before and after the operation using an M9cv ultrasound system (Mindray Medical Solutions, Shenzhen, China) equipped with a L12–4 s transducer (4–12 MHz). Dynamic imaging of color Doppler was obtained in a longitudinal section for at least 30 s under high rate flow mode. The intratesticular blood flow of the testicle was graded on a scale of 0–4 by a sonographer with over 8 years of experience in clinical US diagnosis, based on the criteria established by Coley et al. [35]. This evaluation was conducted without knowledge of the pathological results.

2.4. Photoacoustic imaging system and data acquisition

As illustrated in Fig. 1, to collect the PA signal, we used a 128-element concave ultrasound transducer array (Japan Probe Corporation, Yokohama, Japan) as described in our previous research [36], [37], [38]. The array was connected to homemade pre-amplifiers with a gain of 56 dB and eight DAQ modules (5105, NI). The transducer had a central frequency of 5.0 MHz and a radius of 50 mm. We employed a Q-switched Nd:YAG pumped optical parametric oscillator system (Surelite, Continuum, California) for illumination in the range of 700–960 nm. The pulse duration was approximately 4 ns at a repetition rate of 20 Hz. The excitation laser was transmitted to the tissue through an optic fiber bundle, providing line-shaped illumination with dimensions of 40 mm (long axis) × 10 mm (short axis).

Fig. 1.

Schematic of the PA imaging system.

The optical fluence incident on the tissue surface was estimated to be 5 mJ/cm², which is below the maximum permissible exposure limit recommended by the American National Standards Institute [26]. To account for fluctuations in the laser output, 5% of the light energy was directed through a beam splitter to a laser energy meter (Pulsar-1, Ophir, Jerusalem, Israel).

During the experiment, the PA signals captured by the transducer underwent amplification through a pre-amplifier consisting of 64 channels with a bandwidth ranging from 0.2 to 10 MHz and a gain of 56 dB. The signals were then subjected to a 2:1 multiplexing and averaged ten times under a laser working frequency of 10 Hz or less. Subsequently, the signals were transferred to a 64-channel analog-to-digital system comprising eight PXIe5105 cards (National Instrument, Texas, USA). The PAI system was controlled by an onboard computer (PXIe8840, National Instrument, Texas, USA) which also functioned as a device for signal storage. Each PAI image was acquired within approximately 1 s. The PA image reconstruction was performed using a delay-and-sum algorithm in Matlab (R2016b, Mathworks, Inc., MA, USA), as previously mentioned [39].

The PAI was assessed before and at 2, 4, and 6 h after TT surgery. The animals were placed in a supine position on an oblique panel in a sink, with the water temperature maintained at around 35 ℃. The PAI probe was positioned longitudinally directly above the testicle and illuminated vertically. The PA scanning was performed in the largest longitudinal section, and signals were consistently acquired at wavelengths of 760 nm and 840 nm, with an acquisition rate of 1 frame per second. The data were averaged over time for 10 consecutive frames.

2.5. Image reconstruction and analysis

All data were reconstructed in real-time using the back-projection algorithm in Labview (NI, USA) and stored for subsequent processing by Matlab. To eliminate noise, a 0.2–10 MHz filter was applied during offline processing. Additionally, a laser energy meter (Pulsar-1, Ophir, Jerusalem, Israel) was used to eliminate the impact of optical fluence variation at the two wavelengths (840 nm and 760 nm) on sO₂ calculation, as shown in Fig. 1.

In this study, it was assumed that HbR and HbO₂ are the primary absorbers.

Then, the generated PA signal intensity ($P\left(\stackrel{\rightharpoonup }{r},\lambda \right)$) can be described as follows:

$$P\left(\stackrel{\rightharpoonup }{r},\lambda \right)=\Gamma {\mu }_a\left(\stackrel{\rightharpoonup }{r},\lambda \right)\Phi \left(\stackrel{\rightharpoonup }{r},\lambda \right)$$

Where $\Gamma$ is known as the Grueneisen coefficient, ${\mu }_a\left(\stackrel{\rightharpoonup }{r},\lambda \right)$ and $\Phi \left(\stackrel{\rightharpoonup }{r},\lambda \right)$ are the optical absorption coefficient and optical fluence, respectively. In addition, based on the dual-wavelength acquisitions at 840 nm and 760 nm, the PA data were unmixed to get the concentrations of HbR and HbO₂ according to the Eq. [40].

$${\mu }_a\left(\stackrel{\rightharpoonup }{r},\lambda \right)={\epsilon }_{\mathit{HbR}}\left(\lambda \right)\left[\mathit{HbR}\right]\left(\stackrel{\rightharpoonup }{r}\right)+{\epsilon }_{\mathit{Hb}{O}_2}\left(\lambda \right)\left[\mathit{Hb}{O}_2\right]\left(\stackrel{\rightharpoonup }{r}\right)$$

Where ε ($\lambda$) represents the known molar extinction coefficient of HbR or HbO₂ at wavelength $\lambda$.

The sO₂ mapping calculation in the testis was performed using the formula sO₂ = [HbO₂]/ ([HbO₂] + [HbR]). The imaging of sO₂ distribution allowed for the selection of the region of interest (ROI), which was limited to the 2 mm diameter area located directly beneath the testicular membrane. The average sO₂ per pixel (($\stackrel{\circledR }{s{O}_2}$) was then obtained from the selected ROI (Fig. 2). Three distinct ROIs were identified, and the mean $\stackrel{\circledR }{s{O}_2}$ was subsequently calculated as the final outcome.

Fig. 2.

The calculation of the average oxygen saturation (sO₂) was based on the dual-wavelength PA imaging at 840 nm and 760 nm. The white arrows were used to indicate the border of the testicles, while the white circles denoted the regions of interest (ROI).

2.6. Histopathological assessment

The twisted testes were surgically removed at specific time intervals. The excised tissues underwent histological examination using hematoxylin and eosin (H&E) staining and immunohistochemical (IHC) staining with HIF-1α. Microscopic images of the stained tissues were captured at magnifications of 200 × and 400 × using a light microscope.

Based on the histopathological changes, the twisted testicle was evaluated using Cosentino's histological grading criteria [41], [42]. It was classified as normal (grade 1), ischemia/hypoxia injury (grade 2 or above), or severe injury/necrosis (grade 3 and grade 4). The degree of HIF-1α expression, which indicates the presence of hypoxic conditions, was assessed for each histological section using a scoring system ranging from 1 to 3. The scoring was based on both the expression rate and staining intensity. Cases with no or mild staining and an expression rate below 30% were assigned a score of 1, while those with moderate staining and an expression rate ranging from 30% to 70% were given a score of 2. Finally, cases with intense staining and an expression rate above 70% were scored as 3.

2.7. Statistical analysis

The Shapiro–Wilk test was utilized to analyze the data distribution. One-way ANOVA with Tukey’s post hoc test was used to compare PAI measurements of sO₂ among different groups at various time points for normally distributed data, while non-normally distributed data was analyzed using the nonparametric Kruskal-Wallis test. Spearman's rank coefficients were calculated to evaluate the associations between CDFI score or sO₂ and pathological hypoxic conditions (HIF-1α expressions). The diagnostic performance of the CDFI score and sO₂ in TT diagnosis, specifically in identifying pathological ischemia/hypoxia injury or severe injury/necrosis, was assessed by calculating the areas under the receiver operating characteristic (ROC) curves (AUCs) with 95% confidence intervals (CIs). The sensitivity and specificity were determined using the optimal cutoff points that maximized the Youden index. SPSS (version 26.0, SPSS, Chicago, IL, United States), GraphPad software (version 7.00; GraphPad Software, San Diego, CA, USA), or MedCalc (version 10.4, MedCalc Software, Mariakerke, Belgium) were used for statistical analyses and graphs. Differences were considered significant when P values were < 0.05.

3. Results

3.1. The PA and CDFI imaging of testicles

The figures in Fig. 3 depict the pre- and postoperative PA and CDFI images of the testicles. As the degree of torsion increased, the CDFI signals were scarcely observed in groups B and C. The PA imaging displayed the borders of the testicles and the boundary between the scrotal skin and testicular parenchyma. However, no significant correlations were found between the PA signals in the testicular parenchyma and the degree of torsion.

Fig. 3.

The PA imaging at a wavelength of 760 nm (upper, a₁–d₁) and CDFI imaging (down, a₂–d₂) of testicles in each group at 6 h after operation. (a) group S, (b) group A, (c) group B, and (d) group C.

3.2. The changes of sO₂ obtained by PAI

Two hours after torsion, the $\stackrel{\circledR }{s{O}_2}$ of the testes in groups B and C were significantly lower than those in group S. However, no statistically significant difference was observed between groups A and S. At 4 and 6 h postoperatively, the $\stackrel{\circledR }{s{O}_2}$ levels of the testes in the TT groups (groups A, B, and C) were significantly decreased compared to those in group S. However, no significant differences in $\stackrel{\circledR }{s{O}_2}$ levels were detected between the torsion groups at each time point. For each torsion group, the PAI-measured $\stackrel{\circledR }{s{O}_2}$ after surgery was found to be significantly decreased when compared to the preoperative values (as shown in Fig. 4 and Table 1). There were no significant differences in $\stackrel{\circledR }{s{O}_2}$ at different postoperative time points.

Fig. 4.

The changes in mean oxygen saturation were documented by PAI in torsion testes in groups S, A, B, and C. (*: P < 0.05 vs. Pre).

Table 1.

The mean oxygen saturation ($\stackrel{\circledR }{s{O}_2}$) in different degrees of testicular torsion at different time points (%).

Groups	Pre	2 h	4 h	6 h
S	63.29 ± 4.00	62.10 ± 7.07	60.08 ± 2.95	62.16 ± 5.27
A	59.63 ± 2.87	52.69 ± 5.30p	49.43 ± 4.43s,p	48.11 ± 2.53s,p
B	59.85 ± 4.32	50.15 ± 3.31s	39.84 ± 3.16s,a,p	41.35 ± 6.34s,p
C	63.39 ± 5.23	46.31 ± 4.43s,p	46.59 ± 5.83s,p	40.21 ± 8.15s,p

s: P < 0.05 vs. group S; a: P < 0.05 vs. group A;

p: P < 0.05 vs. Pre

The reduction of $\stackrel{\circledR }{s{O}_2}$ (rsO₂) at each time point was determined by comparing it to the preoperative level (Fig. 5). At the 2-hour postoperative mark, both groups B and C exhibited significantly higher rsO₂ values than group S, with measurements of (10.41 ± 1.22)% and (15.05 ± 2.13)%, respectively. Notably, the rsO₂ levels across the experimental groups increased with the duration of torsion time. Until 6 h post-operation, the rsO₂ values in groups B and C reached (20.69 ± 2.01)% and (23.79 ± 6.40)%, respectively. However, no statistically significant differences in rsO₂ were observed among the postoperative time points for the same degree of torsion group.

Fig. 5.

The reduction of oxygen saturation (rsO₂) was documented by PAI at each time point after operation in different groups. (*: P < 0.05 vs. Group S; &: P < 0.05 vs. Group A).

3.3. Histopathology changes

The testicles in group S exhibited normal seminiferous tubules with spermatogenic cells and mature spermatozoa arranged in an orderly manner (Fig. 6a). Out of the 45 animals subjected to experimentation, 80% (36) of the testes showed ischemia/hypoxia injury. The histopathological examination of these testicles revealed varying degrees of interstitial hemorrhage and a reduction in the spermatogenic epithelium (Fig. 6b-d). Additionally, 40% (18) of the testicles experienced severe injury/necrosis, which was evidenced by irreversible changes such as the destruction of bounded seminiferous tubules, sloughed cells in the lumen, and nucleus pyknosis (refer to Fig. 6c, d).

Fig. 6.

Histologic findings of testicles in group S and torsion groups. (a) Testicles in group S. (b) At 4 h post-operation, minor injury with diminished spermatogenic epithelium and mild interstitial hemorrhage were observed in 180° testicular torsion. (c) After 6 h following operation, severe injury with disordered sloughed cells in the lumen (red arrows) and aggravated interstitial hemorrhage were demonstrated in 360° testicular torsion. (d) Necrosis was presented in 720° testicular torsion after 6 h with destructed bounded seminiferous tubules and extensive interstitial hemorrhage. (All images were taken at an original magnification of ×200.).

3.4. Expression of HIF-1α and its correlation with sO₂ obtained by PAI

HIF-1α was scarcely expressed in normal testicular tissue. However, following TT, the expression of HIF-1α increased in both spermatogenic cells and Leydig cells across all experimental groups, as depicted in Fig. 7. Nonetheless, no statistically significant differences were observed in HIF-1α expression among the groups at 2 h post-operation. At 6 h after the operation, HIF-1α expression was significantly higher in groups B and C than in the control group (both P < 0.05), as shown in Fig. 8. In group C, the expression levels of HIF-1α were significantly elevated at 4 and 6 h post-operation compared to pre-operation (P = 0.008 and P = 0.046, respectively).

Fig. 7.

Expression of HIF-1α by immunohistochemistry: (a) There was no significant expression of HIF-1α in normal testicular tissue. (b) The cytoplasm showed moderate staining, and a few positive nuclei were observed in 180°testicular torsion 2 h after the operation. (c) The cytoplasm displayed strong staining, and the number of positive nuclei increased in 360°testicular torsion 2 h after the operation. (d) More spermatogenic cells and Leyding cells with positive nuclei were observed at 4 h after the operation following 720°testicular torsion. (Red arrow: positive nucleus) (Original magnifications ×200).

Fig. 8.

The scores of HIF-1α expression in each group at various times after testicular torsion. (*: P < 0.05 vs. Group S).

At 6 h after the surgery, a negative correlation was observed between the $\stackrel{\circledR }{s{O}_2}$ and HIF-1α expression. The Spearman correlation coefficient was − 0.66 (P = 0.002). Furthermore, a significant positive correlation was found between the expression of HIF-1α and rsO₂, with a correlation coefficient of 0.70 (P = 0.001). However, there was no significant correlation between the CDFI score and HIF-1α expression (P = 0.07).

3.5. The diagnostic performance of PAI in early detecting TT

When using the operation as the benchmark, the measurement of sO₂ by the PA demonstrated exceptional diagnostic capability in detecting TT, as shown by the AUCs of 0.93 (95% CI, 0.82–1.00) for both$\stackrel{\circledR }{s{O}_2}$and rsO₂ (Fig. 9a). Table 2 presents the optimal cutoff values for $\stackrel{\circledR }{s{O}_2}$ and rsO₂, as determined by the Youden index. Referring to these cutoff values, the sensitivity and specificity of $\stackrel{\circledR }{s{O}_2}$ or rsO₂ in the diagnosis of TT were both found to exceed 80%. In contrast, for CDFI, the AUC of scores for diagnosing TT was 0.82 (95% CI, 0.72–0.93), with a sensitivity of 100% and a specificity of 73%.

Fig. 9.

The receiver operating characteristic (ROC) curves of $\stackrel{\circledR }{s{O}_2}$ (represented by blue lines), rsO₂ (represented by red dot lines), and CDFI score (represented by green dot lines) were analyzed for their effectiveness in diagnosing testicular torsion (a), identifying testicular ischemia/hypoxia injury (b), and severe injury/necrosis (c).

Table 2.

The diagnostic capability of oxygen saturation (sO₂) obtained by PAI in detecting testicular torsion.

Variable	AUC (95%CI)	Cutoff value（%）		Sensitivity（%）		Specificity（%）
$\stackrel{\circledR }{s{O}_2}$	0.93 (0.82–1.00）	56.1		100	86.7
rsO2	0.93 (0.82–1.00）		7.35	80	100

3.6. The diagnostic performance of PAI in identifying ischemia/ hypoxia injury after TT

Regarding the pathological changes that occur after TT, the AUCs of $\stackrel{\circledR }{s{O}_2}$ and rsO₂ for identifying ischemia/hypoxia injury were found to be 0.88 and 0.90, respectively. The corresponding sensitivity and specificity were both above 79%, as shown in Fig. 9b and Table 3. Additionally, the AUC of CDFI scores was determined to be 0.81 (95% CI, 0.71–0.92), with a sensitivity of 81.5% and specificity of 72.6%.

Table 3.

The diagnostic capability of oxygen saturation (sO₂) obtained by PAI in identifying the ischemia/ hypoxia injury after testicular torsion.

Variable	AUC (95%CI)	Cutoff value (%)	Sensitivity (%)	Specificity (%)
$\stackrel{\circledR }{s{O}_2}$	0.88 (0.79–0.97)	50.9	79.2	88.9
rsO2	0.90 (0.81–0.98)	8.9	91.7	79.2

3.7. Diagnostic performance of PAI in identifying severe injury/necrosis after TT

The AUCs of $\stackrel{\circledR }{s{O}_2}$ and rsO₂ were both greater than 70% when distinguishing between testicles that had suffered severe injury/necrosis and those that had not (Fig. 9c, Table 4). A diagnostic specificity of 94.4% was found for $\stackrel{\circledR }{s{O}_2}$ using a cutoff value of 53.3%. In distinguishing severe injury/necrosis, the sensitivity was 100% when rsO₂ was above 7.82%. Comparable diagnostic performance was observed for CDFI scores in the diagnosis of severe injury/necrosis, with an AUC of 0.72 (95% CI, 0.59–0.84). The corresponding sensitivity and specificity were 64.3% and 82.1%, respectively.

Table 4.

The diagnostic capability of oxygen saturation (sO₂) obtained by PAI in identifying severe injury/necrosis after testicular torsion.

Variable	AUC (95%CI)	Cutoff value (%)	Sensitivity (%)	Specificity (%)
$\stackrel{\circledR }{s{O}_2}$	0.72 (0.59–0.85）	53.3	45.2	94.4
rsO2	0.74 (0.62–0.87)	7.82	100	47.6

4. Discussion

Timely diagnosis and appropriate treatment of TT are crucial in preventing irreversible testicular injury and ensuring testicular survival [43], [44]. In clinical practice, the decisions to perform detorsion or orchidectomy depends on the duration of symptoms and the surgeon's subjective evaluation of the twisted testes after incision, without prior knowledge of the underlying pathological changes. In this study, we aimed to investigate the potential of PAI in diagnosing TT and determining the degree of testicular injury. To achieve this, we established various degrees of TT models in rabbits and used pathological changes as the gold standard for assessment.

The pathological changes that occur following TT with a rotated vascular pedicle are due to obstructed arterial inflow and venous drainage, resulting in ischemia and hypoxia. This leads to increased capillary pressure and permeability, allowing for the leakage of red blood cells through the damaged basement membrane and widened endothelial cell gap, resulting in varying degrees of interstitial hemorrhage (Fig. 6). Interestingly, in some severe torsion cases, we observed an increase in spotted PA signals. However, this phenomenon was not present in every twisted case, and the increased signals did not consistently correspond with the degree of hemorrhage. This can be attributed to the complex pathological changes that occur after TT. With prolonged duration, venous stasis can lead to microthrombus formation and eventual infarction, which may diminish the PA signal. Unlike CDFI, the PA signals captured are primarily dependent on the concentration of hemoglobin rather than the flow of blood. Thus, finding the rules governing the variations in PA signals after TT, which affects both inflow and outflow, presents a significant challenge. This is one of the reasons we chose to concentrate on dual-wavelength PA functional imaging at 840 nm and 760 nm, as opposed to vascular structural imaging at a wavelength of 534 nm, to evaluate TT.

In cases of ischemia/hypoxia, the testicles are inevitably accompanied by changes in sO₂. Early and accurate detection of hypoxic conditions (sO₂) holds theoretical value in diagnosing TT and assessing injury degree. In recent years, local blood sO₂ detection has been applied in medical fields, such as evaluating tumor malignancy [45], diagnosing cerebral desaturation and vasospasm in stroke patients [46], and predicting the survival rate of plastic surgery flaps [47]. However, limited research has analyzed the significance of sO₂ in TT assessment. Currently, clinical techniques utilized for detecting sO₂ include pulse oximetry, blood oxygen level-dependent magnetic resonance (BOLD-MRI), and positron emission tomography (PET) [48], [49]. Pulse oximetry is predominantly employed for monitoring peripheral blood sO₂, while BOLD-MRI and PET are costly and infrequently utilized diagnostic tools in emergency settings. Near-infrared spectroscopy (NIRS) analysis and diffuse optical tomography (DOT) also employ near-infrared wavelengths ranging from 690 nm to 850 nm to determine the sO₂ in the region of interest, similar to PAI. This is based on the differences in the infrared absorption coefficient of HbO₂ and HbR in biological tissues at different wavelengths [50], [51]. However, the low spatial resolution of both NIRS and DOT presents challenges in imaging the sO₂ distribution of tissues.

PAI is a non-invasive method that quantifies oxygenation and provides spatially resolved sO₂ values for specific tissue regions. Although it is theoretically possible to compute both the total blood oxygen (HbT = [HbO₂] + [HbR]) and sO₂ concurrently, the heterogeneous tissue’s varying light attenuation results in a greater margin of error in the HbT obtained [52]. Moreover, clinicians are more familiar with and commonly use sO₂ instead of the concentration of HbO₂ or HbR. Therefore, we selected sO₂ as the assessment parameter instead of the hemoglobin concentration. It is worth noting that the sO₂ values obtained through PAI are calculated as an average of both arterial and venous values, resulting in a final $\stackrel{\circledR }{s{O}_2}$ value that is the pixel-averaged sO₂ of the entire vascular bed within the ROI. In this study, we successfully demonstrated the use of PAI in measuring changes in sO₂ levels in testicular tissues after TT.

Compared to organs such as the liver and kidney, which have ample blood supply, the testicle exhibits relative hypovascularity. Using PAI, we determined the $\stackrel{\circledR }{s{O}_2}$ of a normal testicle to be (63.29 ± 4.00)%, which is consistent with the previously reported range of 57–61% measured by NIRS [53], [54], [55], [56]. After TT, we observed a significant reduction in the sO₂ levels of the twisted testicle at 2 h. At this time point, group B and C exhibited rsO₂ levels of (10.41 ± 1.22)% and (15.05 ± 2.13)%, respectively, which were significantly higher than those in the control group. Furthermore, the $\stackrel{\circledR }{s{O}_2}$ levels in all torsion groups decreased below 50%. The rsO₂ in groups B and C were (20.69 ± 2.01)% and (23.79 ± 6.40)%, respectively. These findings suggest that PAI is sensitive to hypoxic conditions following TT.

However, group A did not show a significant decrease in rsO₂ compared to the control group at any point during the observation period following TT. This lack of significant decrease in rsO₂ can be attributed to the mild degree of torsion, which did not result in significant injury during the observation period. This finding may also explain the clinical observation that testicles can be salvaged following TT that lasts longer than 6 h. On the other hand, there were no statistically significant variations in rsO₂ levels observed among the experimental groups during the 4-hour and 24-hour postoperative periods. This could potentially be attributed to the diminishing differences in sO₂ levels between varying degrees of torsion as the duration of the experiment progressed.

AydogduO et al. [53] used NIRS to assess changes in sO₂ levels in rats subjected to varying degrees of TT. One hour post-torsion, the authors observed decrease in testicular sO₂ levels from 59% to 41% in rats subjected to 360° TT (representing a reduction of approximately 18%), and from 61% to 36% in rats subjected to 720° TT (representing a reduction of approximately 25%). Three hours post-torsion, the testicular sO₂ levels in the 360° and 720° TT groups decreased by 25% and 30%, respectively. Hallacoglu et al. [55] conducted a study involving four adult rabbits, in which they utilized NIRS to detect changes in testicular tissue sO₂ following 540° and 720° torsion. Their findings indicated a decrease in sO₂ by (36 ± 2)% on the torsion side compared to the opposite side. In contrast to previous reports obtained through NIRS, our investigation measured lower sO₂ using PAI after TT. This discrepancy may be explained by the limitations of NIRS, which has low spatial resolution and may not accurately sample the ROI. The utilization of ROI selection with low spatial resolution over large areas may lead to a decrease in the averaged sO₂ due to unwanted background signals.

HIF-1 is a transcription factor that facilitates adaptive responses to ischemia/hypoxia. HIF-1α, a subtype of HIF-1, is known to increase under hypoxic conditions, and HIF-1α staining is commonly used to reveal pathological hypoxia [57]. Our study observed an increase in HIF-1α expression across all torsion groups, with significant elevation at 6 h post-surgery. Additionally, the sO₂ levels obtained through PAI exhibited significant correlations with the expressions of HIF-1α. These findings suggest that PAI has the potential to be a valuable tool for quantifying tissue oxygenation saturation.

In clinical settings, color Doppler US has been reported to have a sensitivity range of 69.2–100% and a specificity range of 52.9–100% in the diagnosis of TT [9], [58], [59], [60]. It is important to note that the accuracy of US diagnosis depends on the operator, devices used, and characteristics of the population being investigated. Therefore, surgical exploration is strongly recommended when there is clinical suspicion of TT, especially in pediatric patients, regardless of the US examination results [61]. In this experiment, the use of PAI to measure sO₂ showed exceptional ability in the early detection of TT, with AUC values of 93%, surpassing those obtained with CDFI. The corresponding sensitivity and specificity were both above 80%. These excellent diagnostic results suggest that PAI has promising clinical potential for the diagnosis of TT, allowing for timely clinical intervention while avoiding unnecessary surgical exploration.

In situations where the degree and duration of torsion are unknown, PAI demonstrates exceptional diagnostic performance in identifying the presence of pathological ischemic/hypoxic injury in a twisted testicle. The AUCs of $\stackrel{\circledR }{s{O}_2}$ and rsO₂ were 88% and 90%, respectively, surpassing that of CDFI. The corresponding diagnostic sensitivity and specificity were both above 79%. In discriminating whether the testicle had suffered severe injury/necrosis, we observed good diagnostic capabilities in the $\stackrel{\circledR }{s{O}_2}$ and rsO₂ measured by PAI with AUCs of 0.72 and 0.74, respectively. These AUC values were comparable to those of CDFI. When $\stackrel{\circledR }{s{O}_2}$ decreased below 53.3%, the specificity in distinguishing irreversible necrosis increased to 94.4%. This indicates the possibility of unsalvageable TT, and therefore, orchiectomy may be recommended as the preferred treatment. Conversely, using a cutoff value of 7.82% for rsO₂, the sensitivity in identifying severe injury/necrosis reached 100%, while the specificity was only 47.6%. In this situation, the degree of injury may be overestimated. Hence, in the identification of necrosis following TT, PAI exhibits some value, though not ideal. The possible explanation is that in cases of severe injury/necrosis, infarction may occur, and the presence of necrotic tissue may impact the accurate detection of sO₂. Moreover, in certain severe cases of TT, the swollen and thickened scrotal walls, along with heightened vascularity, may augment the absorption and scattering of light. This could potentially weaken the intensity of the light signal that penetrates the testicular parenchyma, hindering detection and resulting in measurement inaccuracies. Therefore, additional research is required to examine the diagnostic potential of photoacoustic sO₂ in assessing severe injury/necrosis after TT.

Although the results presented are promising, it is important to acknowledge several limitations of this study. Firstly, there is currently no universally accepted method for measuring tissue oxygenation saturation, and further verification is necessary to determine whether the sO₂ obtained by PAI accurately reflects actual tissue sO₂. Secondly, while the PAI system utilized in this trial has a strong experimental foundation, this is the first time it has been applied to TT assessment. To achieve functional imaging of the testicles, we optimized the parameters at the expense of temporal resolution, which could potentially lead to sampling errors. Additionally, the algorithm currently employed does not account for additional optical absorption and scattering caused by the scrotal skin and other heterogeneous tissues, which may result in measurement inaccuracies due to the uncertainty of optical fluence distribution. Optical methods will be employed in the subsequent study for reference purposes, with a focus on achieving greater precision in the detection of sO₂. Finally, it is important to note that this is only a preliminary animal study, and further validation of the results is necessary in larger animals and even human volunteers.

Despite the limitations inherent in this study, it represents the first attempt to examine PAI in TT assessment. In summary, the PAI obtained oxygenation saturation at wavelengths of 760 nm and 840 nm demonstrated good diagnostic capabilities for both TT diagnosis and testicular injury assessment. These findings show significant promise for further clinical application.

Funding

This work was supported by Post-Doctor Research Project, West China Hospital, Sichuan University Grant No.2021HXBH039.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Biographies

Qianru Yang is currently a doctor in Department of Ultrasound, West China Hospital, Sichuan University. She received his Ph.D. degree in 2020 from Sichuan University. She is now a post-doctoral fellowship at Sichuan University. Her research mainly focuses on the breast ultrasonography and the clinical application of photoacoustic imaging.

Lulu Yang is an attending physician in West China Hospital of Sichuan University. She got her Ph.D. in Clinical medicine from Sichuan University, China. She has worked in the field of clinical imaging, especially in ultrasonography.

Chihan Peng is currently a doctor in Department of Ultrasound, West China Hospital, Sichuan University. He received his Ph.D. degree in 2019 from Sichuan University. He completed his post-doctoral fellowship in 2022. His research mainly focuses on the liver ultrasonography and interventional ultrasound.

Xiaoxia Zhu is a research assistant of West China Hospital of Sichuan University of China. She received her Master's degree in 2012 from southwest University for nationalities of China with expertise in Veterinary Medicine.

Zhenru Wu is a Laboratory technician of West China Hospital, Sichuan University and Institute of Clinical Pathology at West China Hospital, Sichuan University. She received her master's degree in 2012 from Three Gorges University of China with expertise in ecology.

Lin Huang received the Ph.D. degree in optics from the University of Electronic Science and Technology, Chengdu, China, in 2015. He is currently an Associate Professor with the University of Electronic Science and Technology of China. His research is focused on discovering and developing fundamentally electromagnetic related imaging technologies for in vivo visualization of tissue at both the macroscopic and microscopic scales.

Yan Luo is currently the director and a Professor of Department of Ultrasound, West China Hospital, Sichuan University. She has been engaged in clinical and scientific research in ultrasound. Her current research interests include ultrasonography for abdominal diseases (especially liver diseases) and emerging imaging technologies.

Data Availability

Data will be made available on request.