Value for combination of T₁WI star-VIBE with TWIST-VIBE dynamic contrast-enhanced MRI in distinguishing lung nodules

Abstract

Objective

With the increasing detection rate of lung nodules, the qualitative problem of lung nodules has become one of the key clinical issues. This study aims to evaluate the value of combining dynamic contrast-enhanced (DCE) MRI based on time-resolved imaging with interleaved stochastic trajectories-volume interpolated breath hold examination (TWIST-VIBE) with T₁ weighted free-breathing star-volumetric interpolated breath hold examination (T₁WI star-VIBE) in identifying benign and malignant lung nodules.

Methods

We retrospectively analyzed 79 adults with undetermined lung nodules before the operation. All nodules of patients included were classified into malignant nodules (n=58) and benign nodules (n=26) based on final diagnosis. The unenhanced T₁WI-VIBE, the contrast-enhanced T₁WI star-VIBE, and the DCE curve based on TWIST-VIBE were performed. The corresponding qualitative [wash-in time, wash-out time, time to peak (TTP), arrival time (AT), positive enhancement integral (PEI)] and quantitative parameters [volume transfer constant (Ktrans), interstitium-to-plasma rate constant (Kep), and fractional extracellular space volume (Ve)] were evaluated. Besides, the diagnostic efficacy (sensitivity and specificity) of enhanced CT and MRI were compared.

Results

There were significant differences in unenhanced T₁WI-VIBE hypo-intensity, and type of A, B, C DCE curve type between benign and malignant lung nodules (all P<0.001). Pulmonary malignant nodules had a shorter wash-out time than benign nodules (P=0.001), and the differences of the remaining parameters were not statistically significant (all P>0.05). After T₁WI star-VIBE contrast-enhanced MRI, the image quality was further improved. Compared with enhanced CT scan, the sensitivity (82.76% vs 80.50%) and the specificity (69.23% vs 57.10%) based on MRI were higher than that of CT (both P<0.001).

Conclusion

T₁WI star-VIBE and dynamic contrast-enhanced MRI based on TWIST-VIBE were helpful to improve the image resolution and provide more information for clinical differentiation between benign and malignant lung nodules.

As a symbol of early lung cancer, accurate management of lung nodules is the focus of clinical attention, especially the identification of early lung cancer^[1-2]. The 10-year survival rate of patients with stage I lung cancer after surgical treatment can reach 92%^[3], therefore, how to accurately distinguish benign and malignant lung nodules in the early stage becomes crucial. Chest CT is the most commonly used non-invasive imaging examination, but the radiation problems caused by multiple and repeated scans cannot be ignored^[4]. It is estimated that if annual low-dose computed tomography (LDCT) screening is carried out among 50-75 years old, including current or former smokers, the malignant tumors caused by medical radiation will reach 0.5%-5.5%^[5]. Besides, it is difficult to differentiate malignant tumor from atypical granuloma when patients lacked typical clinical symptoms, and it has been reported^[6] that about 25% of tuberculous granulomas were initially diagnosed as lung cancer because of overlapping signs between these two diseases, leading to unnecessary surgical treatment.

Different from CT, MRI allows multiple follow-ups because of its non-radiation, which is more beneficial for high-risk patients who need multiple and close follow-ups^[7-9]. A previous study^[10] showed that conventional MRI could essentially distinguish mass pulmonary tuberculosis from lung cancer, but the role of differentiation between benign and malignant pulmonary nodular lesions remains to be further explored. Dynamic contrast enhanced-MRI (DCE-MRI) is a non-invasive, non-radiation functional magnetic resonance imaging method^[11]. Through multi-phase dynamic scanning after slow injection of gadolinium, DCE-MRI can ideally reflect the changes in hemodynamics and blood perfusion inside the lesions. At present, it has been widely used in the evaluation of breast cancer, prostate cancer, liver tumors, and other tumors^[12-13]. While based on conventional sequences, DCE-MRI images are acquired with thick layer thickness and low signal-to-noise ratio only in the different breath-holding states, meaning that the image quality is affected by breathing state and maybe existing movement to some extent, so it isn’t suitable for patients with uncooperative breathing activity or poor general condition.

Recently, DCE-MRI based on time-resolved imaging with interleaved stochastic trajectories-volume interpolated breath hold examination (TWIST-VIBE) has been developed fast because it can offer both high temporal and spatial resolution simultaneously. It is considered a potential candidate for conventional DCE-MRI, demonstrating promising results in breast disease^[14-17]. Meanwhile, T₁-weighted free-breathing star-volumetric interpolated breath hold examination (T₁WI star-VIBE) is increasingly used because it can breathe freely and can improve image quality. Star-VIBE is a sequence that can scan the whole lung in the state of free breathing, which has more advantages than other conventional sequences. First of all, when scanning in the state of free breathing, the patient’s tolerance and comfort can be further improved. Secondly, because its main principle is oversampling in K space, it also has a excellent averaging effect on motion and heartbeat artifacts. In addition, the image obtained by the star-VIBE sequence has a thinner layer thickness of about 1 mm, a higher signal-to-noise ratio, and better display clarity for small lesions, all of which are helpful in defining the regions of interest (ROI) of small lesions and enabling the collection of more precise qualitative and quantitative data.

This study aims to evaluate the signal characteristics from unenhanced T₁-weighted volumetric interpolated breath hold examination (T₁WI-VIBE) and contrast-enhanced T₁WI star-VIBE, the curve characteristics from DCE-MRI scanning, and their related qualitative and quantitative qualitative parameters, to determine whether the above indicators can differentiate malignant from benign lung nodules.

1. Materials and methods

1.1. Clinical data and baseline characteristics

A total of 79 patients from June 2019 to September 2021 were retrospectively included. The inclusion criteria were as follows: 1) lung nodules larger than 5 mm; 2) without contraindication for MR imaging, including ferromagnetic implants, pacemaker, and claustrophobia. All the patients with malignant tumors were confirmed by post-operation pathology or puncture biopsy. The included 7 tuberculosis cases (9 nodules) meet the clinical diagnostic criteria for tuberculosis. Moreover, tuberculosis clinical diagnostic criteria include^[18]: 1) a history of active pulmonary tuberculosis exposure; 2) the location and morphological characteristics of tuberculosis were consistent with the imaging findings; 3) immunological tests were positive; 4) After anti-tuberculosis treatment, the lesion was reduced and the symptoms were relieved. Other benign lesions except tuberculosis were confirmed by pathology after surgical resection. Exclusive criteria: 1) unclear image display or severe artifacts affecting diagnosis; 2) those who only performed MRI scans without enhanced CT scans. All nodules included were classified into a malignant group (n=58) and a benign group (n=26) based on final diagnosis. This study was approved by the Medical Ethical Committee of our hospital (Approval number: 2021-175).

1.2. CT scanning

The dual-energy CT (Somatom Force, Siemens Healthcare, Erlangen, Germany) was used for scanning. All participants in this study underwent an enhanced CT scan with an intravenous injection of 1.2 mL/kg of ioversol at a rate of 2.5 mL/s using an injector. The tube voltage was 100 kV, the tube current was 71 mA, the scanning range was from the apex of both lungs to the lower part of the bilateral diaphragm, and the lung window and mediastinal window were collected at the end of inspiration, with the thickness of 1 mm and 5 mm respectively.

1.3. MRI scanning

All MRI examinations were performed on a 3.0 T MR scanner (Magnetom Skyra, Siemens Healthcare, Erlangen, Germany) with an 18-channel system combining a 6-element anterior body matrix coil and a 12-element posterior spine matrix coil. Firstly, unenhanced T₁WI VIBE images at the flip angle of 9° and T₂-weighted half-fourier acquisition single-shot turbo spin echo (T₂WI-HASTE) images were acquired. Then, free-breathing DCE-MRI based on the TWIST-VIBE sequence was performed with an intravenous injection of 0.1 mmoL/kg of gadoteridol injection at a rate of 2.5 mL/s by using a power injector. The TWIST-VIBE sequence continuously scanned 50 times during 325 s. After DCE-MRI scanning, the contrast-enhanced conventional T₁WI-VIBE and T₁WI star-VIBE were performed. The total scanning time was about 14 min, and the main scanning parameters were shown in Table 1.

Table 1.

MRI scanning parameters

Categories	TR/TE	Thickness/mm	Flip angle/(°)	Matrix/mm3	FOV/mm2	Scanning time/s	Respiratory control
T2WI-HASTE	1 400/87	5.0	160	1.25×1.25×6.0	380×380	60	Breathless scanning
T1WI-VIBE/T1WI-VIBE+C	3.93/1.89	3.5	9	1.2×1.2×3.0	380×380	17	Breathless scanning
TWIST-VIBE+C	5.08/1.71	3.5	15	1.67×1.67×3.5	320×320	325	Free breathing
T1WI star-VIBE+C	2.79/1.39	1.2	5	1.2×1.2×1.2	380×380	420	Free breathing

T₂WI-HASTE: T₂-weighted half-fourier acquisition single-shot turbo spin echo; T₁WI-VIBE: T₁-weighted volumetric interpolated breath hold examination; C: Contrast enhanced; TWIST-VIBE: Time-resolved imaging with interleaved stochastic trajectories-volumetric interpolated breath hold examination; T₁WI star-VIBE: T₁-weighted free-breathing star-volumetric interpolated breath hold examination; TR: Tepetition time; TE: Echo time; FOV: Field of view.

1.4. Image analysis

All images are uploaded to our institution’s picture achieving and communication system (PACS). Most MRI image analyses were conducted in the PACS viewer system except the DCE MRI data, which were analyzed by the MR Tissue 4D analysis (Sygno. via, Siemens Healthcare, Erlangen, Germany). According to the previous study about the delineation of ROIs, all ROIs after MRI scanning were carried out on 3 consecutive slices, and the central necrotic part of the nodule, large vessels, and prominent artifacts were avoided as much as possible to ensure the accuracy of the measurement^[19-20]. Two chest radiologists (one with more than 10 years of diagnostic experience and the other with less than 3 years of diagnostic experience) who were blinded to clinical and histological findings analyzed the nodules and reached an agreement by consultation. The CT images of nodules were mainly judged by their type, size, shape, lobulation and burr, pleural retraction, and vacuole sign.

We evaluated 5 aspects of nodules. Firstly, the diameter of the nodule was measured, if the long and short diameters of the nodule were inconsistent, the average length was calculated. Secondly, the signal intensity of the nodule from unenhanced T₁WI -VIBE was evaluated compared with the thoracic medulla and rhomboid muscle: 1) hypo-intensity signal which refers to lower than muscle signal; 2) intermediate signal which refers to between muscle and thoracic medulla signal; 3) hyper-intensity which refer to higher than the signal intensity of the thoracic medulla. SL is the signal intensity of the lesion, and SM is the signal intensity of the rhomboid muscle (or thoracic medulla). To exclude individual differences and the influence of smoking, we calculated the lesion’s signal intensity ratio (SIR) to rhomboid muscle and thoracic medulla on T₁WI-VIBE images: SIR=SL/SM(SRM) $\times$ 100%. Thirdly, taking CT as a reference, we evaluated the contrast-enhanced MRI image quality based on T1WI star-VIBE and T₁WI-VIBE, we mainly adopted subjective evaluation based on image quality, nodule display, and artifacts. A 5-point system was adopted to evaluate the image quality using the following evaluation criteria. Score=1: the image is blurred with obvious artifacts and cannot be used for diagnosis; score=2: the image is blurred, there are obvious artifacts, but it does not affect the diagnosis; score=3: there is a fuzzy boundary with slight artifact on some layers, but the image does not have a problem for diagnosis; score=4: the lesion boundary is slightly blurred and there are minor artifacts, but the image can be used for diagnosis; score=5: the lesion boundary is clear without any artifacts, and the image can be used for diagnosis. Moreover, the nodule’s enhancement uniformity was evaluated, When the lesion signal increased uniformly after enhancement,it means homogeneous, patchy, annular, or irregular reinforcement means heterogeneous. In addition, the DCE curves of the nodule were analyzed by the MR Tissue 4D analysis, and it mainly includes 4 types: fast-rising and falling type (type A, Figure 1A); fast-rising platform type (type B, Figure 1B); slowly rising type (type C, Figure 1C); stationary type (type D). Finally, we further calculated the quantitative parameters, such as wash-in, wash-out time, time to peak, and other parameters such as volume transfer constant (Ktrans), interstitium-to-plasma rate constant (Kep), and fractional extracellular space volume (Ve) based on the toft model. It should be noted that due to the low density of some of the nodules we included, they could not be displayed in the dynamic-enhanced image. Therefore, we only outlined the qualitative and quantitative data of about half of the nodules.

Figure 1. Diagram of type A (A), type B (B), and type C (C) of dynamic enhancement curves.

1.5. Diagnostic performance of MRI and CT

MRI images of nodules were mainly judged by its T₁WI-VIBE unenhanced scan signal, T₁WI star-VIBE contrast enhancement uniformity, DCE curve, and DCE quantitative parameters. Then, taking pathology as the gold standard, the sensitivity and specificity of the contrast enhanced MR and CT methods were calculated and compared respectively.

1.6. Statistical analysis

SPSS Version 21.0 was used for statistical analysis, the quantitative data were tested using the independent sample t-test and Mann-Whitney U test. The qualitative data were tested using the χ² test. P<0.05 was considered statistically significant.

2. Results

2.1. Clinical and baseline data of patients

A total of 79 patients (47 females and 32 males) with 84 lung nodules (26 benign and 58 malignant, 2 patients with 3 nodules and 1 patient with 2 nodules) were retrospectively included, including 45 solid nodules and 39 ground glass nodules. The workflow was shown in Figure 2. The specific clinical and baseline data are shown in Table 2. The diameter of malignant and begin nodule was (12.6±4.7) and (12.0±4.5) mm, respectively. There were no significant differences between in age, gender, nodular diameter, location distribution, and whether adjacent to the subpleural area (all P>0.05, Table 2). The final diagnoses of all nodules were shown in Table 3.

Figure 2. Workflow of this study.

Table 2.

Clinical and baseline data of included patients

Groups	n	Age/years	Diameter/No.
			5-10 mm	>10-15 mm	>15-20 mm	>20 mm
Malignant	58	55.5±10.6	20	20	15	3
Benign	26	52.1±11.8	10	11	3	2
χ 2/t		-1.296	-0.566
P		0.199	0.573

Groups	Nodule location/No.					Relationship to the pleura/No.
	Left upper lobe	Left lower lobe	Right upper lobe	Right middle lobe	Right lower lobe	Subpleural	Non-subpleural
Malignant	16	12	17	2	11	28	30
Benign	6	6	7	2	5	12	14
χ 2/t	0.032					0.032
P	0.857					0.857

Table 3.

Pathologic findings of lung nodules included in this study

Pathological classification	Pathologic findings	Amounts	Pathological classification	Pathologic findings	Amounts
Malignant	Adenocarcinoma	49	Benign	Inflammation (granuloma)	13
	Lymphoma	4		Tuberculoma	9
	Lymphoepitheliomatoid carcinoma	2		Hamartoma	2
	Basal cell carcinoma	1		Sclerosing lveolar cell tumor	1
	Signet-ring cell carcinoma	1		Pulmonary alveolar proteinosis	1
	Carcinoid tumor	1

2.2. Image quality evaluation

As shown in Figure 3, the T₁WI star-VIBE contrast MRI image is more clear than the other two. In the 84 cases, the subjective score over 4 was 100%. The consistency between the scores of the 2 observers was high, and the Kappa value was 0.875 (Table 4).

Figure 3. Comparison of image quality after different MRI scanning sequences.

A: T₁WI-VIBE; B:T₁WI-VIBE+C; C: T₁WI star-VIBE+C. T₁WI-VIBE: T₁-weighted volumetric interpolated breath-hold examination; C: contrast enhanced; T₁WI star-VIBE: T₁-weighted free-breathing star-volumetric interpolated breath-hold examination.

Table 4.

Result of the subjective image quality score

Categories	Score/No.
	1	2	3	4	5
Observer1
T1WI-VIBE	0	2	43	32	7
T1WI-VIBE+C	0	1	5	40	38
T1WI star-VIBE+C	0	0	0	16	68
Observer2
T1WI-VIBE	0	2	41	35	6
T1WI-VIBE+C	0	1	7	35	41
T1WI star-VIBE+C	0	0	0	13	71

T₁WI-VIBE: T₁-weighted volumetric interpolated breath-hold examination; C: Contrast enhanced; T₁WI star-VIBE: T₁-weighted free-breathing star-volumetric interpolated breath-hold examination.

2.3. Difference of unenhanced T₁WI-VIBE signal intensity, T₁WI star-VIBE contrast enhancement uniformity, DCE MRI parameters between benign and malignant nodules

As shown in Table 5, among the malignant lung nodules, 82.8% of nodules showed T₁WI-VIBE hypo-intensity, while only 30.8% among benign nodules (P<0.001). For the contrast-enhanced T₁WI star-VIBE, most of the malignant lung nodules showed homogeneous enhancement and the difference was statistically significant with benign lung nodules (P=0.003). The difference in the DCE curve between benign and malignant lung nodules was statistically significant (P<0.001). Among malignant lung nodules, 53.4% of the nodule DCE were type A, while only 15.4% of benign lung nodules showed a type A DCE curve. The wash-out of malignant lung nodules was significantly lower than that of benign lung nodules (P=0.001); the differences of wash-in, Ktrans, Kep, TTP, AT, PEI, and Ve were not statistically significant between begin and malignant lung nodules (all P>0.05, Table 5).

Table 5.

MRI features between benign and malignant lung nodules

Groups	n	Unenhanced T1WI-VIBE signal/No.			Enhanced uniformity/No.		DCE curve type/No.
		Hypo-intensity	Iso intensity	Hyper-intensity	Homogeneous	Heterogeneous	Type A	Type B	Type C
Malignant	58	48**††††	5	5	42	16	31	27‡	0‡‡‡§§
Benign	26	8	6	12	10	16	4	14	8
χ 2/Z		21.916			8.775		23.402
P		<0.001			0.003		<0.001

Groups	Qualitative data					Quantitative data
	Wash-in/s	Wash-out/s	TTP/s	AT/s	PEI/min-1	Ktrans/min-1	Kep	Ve
Malignant	0.276	-0.008	0.811	0.544	0.279	0.250	0.856	0.270
Benign	0.089	0.003	0.811	0.547	0.238	0.190	0.770	0.266
χ 2/Z	-1.673	3.223	0.604	0.305	-0.582	-0.166	0.364	-0.331
P	0.094	0.001	0.546	0.761	0.561	0.883	0.731	0.756

**P<0.01 vs isointensity; †††P<0.001 vs hyper-intensity; ‡P<0.05, ‡‡‡P<0.001 vs type A; §§P<0.01 vs type B. T₁WI-VIBE: T₁-weighted volumetric interpolated breath hold examination; DCE: Dynamic contract enhanced;TTP: Time to peak; AT: Arrival time; PEI: Positive enhancement integral; Ktrans: Volume transfer constant; Kep: Interstitium-to-plasma rate constant; Ve: Fractional extracellular space volume.

The benign nodules showed heterogeneous enhancement after T₁WI star-VIBE contrast-enhanced MRI, and the DCE curve showed type C (Figure 4).

Figure 4. Physical examination of a 50-year-old man revealed solid nodules in the left upper lung with a long history of smoking.

A: PET-CT scanning. A few burrs were seen on the edge of the solid nodule of the left upper lung, the diameter was 11 mm×8.6 mm, and the glucose metabolism uptake was increased, SUV_max=7.4. B and C: MRI scanning. T₁WI-VIBE unenhanced MRI showed a high signal. The enhancement of the left upper lung nodule was heterogeneous. D: Dynamic enhancement curve was type C. E-F: Two months (E) and 6 months (F) after anti-tuberculosis treatment, the left upper lung nodules were significantly smaller than before.

2.4. Comparison of diagnostic efficacy between CT and MRI

Since we found that T₁WI-VIBE hypo-intensity, homogeneous enhancement and type A curve might indicate malignant lesions of lung nodules, we calculated the sensitivity and specificity based on the above parameters respectively, and finally combined these 3 indicators to calculate the total diagnostic efficacy of MRI. And the detailed information was showed in Table 6, among the 84 lung nodules included in this study, the sensitivity, specificity based on MRI were 82.70%, 69.23%, respectively. Based on enhanced CT, the sensitivity, specificity were 80.50%, 57.10%, respectively. The accuracy of diagnosis based on MRI was higher than that of CT (P<0.001). Besides, 4 significant MRI parameters, including T₁WI-value, T₁WI/R, T₁WI/S and wash-out, were performed ROC test, and their AUC was 0.72, 0.74, 0.73, and 0.77 respectively, which showed a good prediction of lung malignant nodules (Figure 5).

Table 6.

Comparison of diagnostic efficacy between MRI and enhanced CT based on pathological results

Diagnostic performance index	T1WI-VIBE hypo-intensity	Homogeneous enhancement	Type A DCE curve	MRI	CT
P	<0.001	0.013	0.006	<0.001
Sensitivity/%	82.70	72.40	53.40	82.76	80.50
Specificity/%	69.23	61.50	84.60	69.23	57.10
AUC	0.759	0.670	0.690

DCE: Dynamic contrast enhanced; AUC: Area under the curve.

Figure 5. Receiver operating characteristic curve of MRI parameter.

3. Discussion

In this study, the combination of T₁WI star-VIBE and dynamic contrast-enhanced MRI based on TWIST-VIBE can effectively distinguish benign and malignant lung nodules. Our study showed that by using T₁WI star-VIBE contrast-enhanced MRI, the image quality is approved effectively, which can achieve equal high image resolution compared with CT. Besides, the hypo-intensity on unenhanced T₁WI-VIBE, the type A of DCE curve, and short wash-out time mean that lung nodules were more likely to be malignant. Our results have also showed that the accuracy of diagnosis based on MRI is higher than that of CT, which is moderately consistent with pathological diagnosis.

A previous study^[10] showed that conventional MRI can essentially distinguish mass pulmonary tuberculosis from lung cancer, after combining with the characteristics of T₂WI unenhanced signal, and the sensitivity can be up to 96.7%. However, the role of differentiation between benign and malignant pulmonary nodular lesions remains to be further explored. T₁WI-VIBE^[21] is a derivative sequence of gradient echo sequence (GRE), which has a shorter TE value than traditional GRE sequence and can obtain images with fewer artifacts, higher signal-to-noise ratio (SNR), and higher image quality in a short time. Therefore, the analysis of unenhanced T₁WI-VIBE signal intensity will supply clearer images together with more accurate and reliable results in this study. On T₁WI-VIBE unenhanced signal intensity, we discover differences between benign and malignant lung nodules. Furthermore, in malignant lung nodules, the proportion of hypo-intensity on unenhanced T₁WI-VIBE was 82.8%, which may be related to different pathological types of tumors, inconsistent cell water content, and protein content^[22]. In benign lung nodules, the hypo-intensity on T₁WI-VIBE accounted for only 30.7%. Interestingly, the 9 tuberculoma included in this study showed iso-intensity or high intensity on unenhanced T₁WI-VIBE, while after T₁WI star-VIBE contrast-enhanced MRI, most of them showed heterogeneous enhancement. This result may relate to caseous necrosis and lipid deposition, which is consistent with previous studies^[10], but further conclusions still need to be verified by a large sample. To sum up, with star-VIBE, we can obtain a thinner layer and clearer image in the state of free breathing, and our study also confirmed.

DCE-MRI is a non-invasive and non-radiation functional magnetic resonance imaging method^[23]. Through DCE-MRI scanning, the DCE curve, and qualitative and quantitative parameters of the lesion can be obtained. As the vascular maturity, growth rate, and proliferation rate of benign and malignant nodules vary, these indicators can be used to identify them. While conventional DCE-MRI scanning requires high temporal resolution and respiratory control, so it cannot be applied to the lung MRI scanning. In this study, we use DCE based on TWIST-VIBE which is a combination of TWIST and VIBE sequences. In the TWIST acquisition strategy, the k-space periphery is severely under-sampled. While the similarity of adjacent time frames allows the peripheral k-space data of adjacent time points to be shared, resulting in high spatial resolution^[24-25]. After combining with VIBE, which can provides superior fat suppression with complementary double-echo dixon fat and water separation, as well as a high signal-to-noise ratio. As a result, TWIST-VIBE allow high temporal resolution and spatial resolution to develop an extremely fast DCE-MRI protocol. Besides, it could allow free breathing which could provide a comfortable experience for patients. Although it has been investigated in several previous studies ^[25-26], its value in lung MRI and the resulting clinical consequences remain to be determined.

This study has discovered that all malignant nodules displayed type A or type B curves after DCE-MRI scanning based on TWIST-VIBE, which may be related to the lung tumor’s abundant immature angiogenesis, a significant endothelial cell gap, and a significant capillary permeability^[27]. As a result, it has a faster inflow time and outflow time. However, when the patients present with acute inflammation, inflammation leads to the increase of peripheral vascular permeability and blood flow velocity, which may be the reason for no type A or type B curve of benign lung nodules in this study. We further analyze the differences in qualitative and quantitative index between the malignant and benign lung nodules and found that the washout time of malignant nodules were shorter, and the Ktrans and Kep values were larger, which were consistent with the results of qualitative curve. Unfortunately, there is no statistically significant difference between the values of Ktrans and Kep. This may be due to the limited sample size of the study and the fact that fewer patients came to hospital because of the COVID-19 epidemic. Finally, based on the gold standard of pathology, this study has discovered that MRI had higher sensitivity, specificity, and diagnostic accuracy than CT in identifying benign from malignant lung nodules. The properties of MR multi-parameter imaging, which can provide more functional information in addition to the morphology of the lesion, may be related to the diagnostic accuracy of MRI, which is moderately consistent with the gold standard of pathological diagnosis^[28-29]. In the future, we also plan to use faster compressed sensing acceleration technology to perform DCE scanning.

Recently, Dynamic contrast enhanced computed tomography (DCE-CT) is increasingly used in the differentiation of lung nodules because it can also provide blood flow information about lesions^[30]. DCE-CT involves the acquisition of a dynamic series of images of nodules before and after intravenous administration of iodine-based contrast agents which can quantify the enhancement of lung nodules. The enhancement reflects the extent of the vessel and has high sensitivity and moderate specificity for the diagnosis of Solitary lung nodules. A meta-analysis^[31] of 23 studies involving 2 397 patients with lung nodules, and it conclude that DCE-CT has good diagnostic effectiveness with a pooled sensitivity and specificity of 94.8% and 75.5% respectively. But a study^[32] has proved that DCE-CT is helpless in the characterization of pure ground glass nodules (GGO) and part solid nodules with a soft tissue component that measures <8 mm, which is obviously not conducive to the wide application of DCE-CT. In this study, nearly half of nodules less than 10 mm in diameter are included, but we conclude that DCE-MRI still has good diagnostic accuracy, which is superior to DCE-CT. Moreover, at present, the number of scanning periods of most DCE-CT is significantly smaller than that of DCE-MRI in our study, and the interval between the scanning periods is long, which is not conducive to microcirculation perfusion analysis of the lesions. Last but not least, the radiation problem of DCE-CT cannot be ignored. In the future, more high-quality multicenter studies are needed to confirm the diagnostic value of DCE-CT.

This study has some limitations: 1) our study is a retrospective study, which may cause selection bias. 2) the sample size included in this study is relatively small, especially the number of benign lung nodules. Therefore, the accuracy of the conclusion needs to be verified in a larger sample. 3) since most of the nodules were ground glass nodules, some quantitative analysis of nodules was not analyzed. In the future, we will use large samples, incorporating machine learning and other research methods to further explore.

In conclusion, with strong sensitivity, specificity, and reasonable consistency, the combination of T₁WI star-VIBE and DCE MRI based on TWIST-VIBE may effectively distinguish between benign and malignant lung nodules. As a excellent diagnostic method, MR can be used independently in distinguishing benign and malignant lung nodules, which is expected to improve diagnostic accuracy and reduce the radiation dose to patients.

Funding Statement

This work was supported by the Program of Clinical Research Center for Medical Imaging in Hunan Province (2020SK4001), the New Medical technology of the Second Xiangya Hospital of Central South University ([2019]04 and [2021]175), and the Extracurricular Scientific Research Training Program for Medical Students of Central South University (202029KT519), China.

Conflict of Interest

The authors declare that they have no conflicts of interest to disclose.

AUTHORS’CONTRIBUTIONS

HU Junjiao and LIU Meitao Data curation, formal analysis and original draft writing; ZHAO Wei and DING Ziyan Investigation and paper modification; WU Fang and HU Wen Data collecting and analysis; GUO Hu and ZHANG Huiting Image post-processing; HU Pei, LI Yiyang, OU Minjie, and HAN Danqi Investigation; CHEN Xiangyu Funding acquisition, supervision and paper modification.The final version of the manuscript has been approved and read by all authors. All authors have read and agreed to the final text.

Note

http://xbyxb.csu.edu.cn/xbwk/fileup/PDF/202304581.pdf