Development and validation of a combined model based on dual-sequence MRI radiomics for predicting the efficacy of high-intensity focused ultrasound ablation for hysteromyoma

Abstract

Objectives

To determine the value of dual-sequence magnetic resonance imaging (MRI)-based radiomics in predicting the efficacy of high-intensity focused ultrasound (HIFU) ablation for hysteromyoma.

Methods

A total of 142 patients with 172 hysteromyomas (95 hysteromyomas from the sufficient ablation group, and 77 hysteromyomas from the insufficient ablation group) were enrolled in the study. The clinical–radiological model was constructed with independent clinical–radiological risk factors, the radiomics model was constructed based on the optimal radiomics features of hysteromyoma from dual sequences, and the two groups of features were incorporated to construct the combined model. A fivefold cross validation procedure was adopted to validate these models. A nomogram was constructed, applying the combined model in the training cohort. The models were assessed with receiver operating characteristic (ROC) curves and integrated discrimination improvement (IDI). An independent test cohort comprising 40 patients was used to evaluate the performance of the optimal model.

Results

Among the three models, the average areas under the ROC curves (AUC) of the radiomics model and combined model were 0.803 (95% confidence interval (CI): 0.726–0.881) and 0.841 (95% CI: 0.772–0.909), which were better than the clinical–radiological model in the training cohort. The IDI showed that the combined model had the best prediction accuracy. The combined model also showed good discrimination in both the validation cohort (AUC = 0.834) and the independent test cohort (AUC = 0.801).

Conclusion

The combined model based on the dual-sequence MRI radiomics is the most promising tool from our study to assist clinicians in predicting HIFU ablation efficacy.

Keywords:

• Hysteromyoma
• high-intensity focused ultrasound
• radiomics
• magnetic resonance imaging
• prediction

Introduction

Hysteromyoma is a common benign tumor in women of reproductive age, with an incidence rate of more than 70% [1]. Myoma originates from uterine smooth muscle tissue, and its growth depends on estrogen and progesterone [2]. Twenty to 50% of patients have corresponding symptoms, such as menstrual disorders, abdominal pain and loss of fertility, which seriously reduce the quality of life of patients [3,4].

Common treatment options for hysteromyoma include hysterectomy, drug therapy and minimally invasive surgery [5]. Even though hysterectomy remains a recognized treatment [6], it has certain limitations. For example, the patient age must be considered when determining whether to perform this procedure. For patients who want to retain fertility, the radical surgery should be avoided. Due to the lack of sufficiently large sample sizes and high-quality studies, uncertainty persists in drug treatment and the possibility of adverse reactions after drug withdrawal [4]. In minimally invasive surgery, myomectomy is currently the gold standard for the treatment of hysteromyoma in women who desire pregnancy. However, this surgery has a long convalescence. High-intensity focused ultrasound (HIFU) has the potential to become an alternative to myomectomy [7]. The HIFU principle is to ablate the target tissue under the guidance of monitoring images by relying on the thermal and mechanical energy of ultrasonic waves [8]. The purpose of this method to cause cell coagulation necrosis only at the focus within the target volume and minimize the damage to the surrounding and covered tissues [9]. Numerous clinical trials have confirmed that HIFU has the advantages of safety, effectiveness and rapid recovery after treatment, especially for the patients who want to preserve the uterus or wish to remain fertile [7,10,11]. The non-perfusion volume ratio (NPVR) is an important means to evaluate ablation efficacy. NPVR was associated with reduced myoma volume and improvement of symptoms during medium-long term follow-up [12]. Therefore, the preoperative prediction of NPVR has great significance for accurately selecting patients and improving the success rate of HIFU treatment.

Previous studies have found that when myoma presents high T2-weighted imaging (T2) signal intensity and a high apparent diffusion coefficient (ADC) value, the NPVR is poor [13,14]. However, these traditional imaging parameters can be disturbed by many factors. Even for hysteromyoma with the same signal before operation, great differences in the NPVR persist after operation. As an imaging-based technology, radiomics uses computer algorithms to extract and analyze many quantitative features that cannot be recognized by the naked eye and provides spatial coding information about normal and abnormal tissue structure and function, which can be used in clinical decision support systems to improve the accuracy of diagnosis and prognosis [15–17]. Recently, some studies have assessed the value of radiomics models based on nonenhanced magnetic resonance imaging (MRI) in predicting the efficacy of HIFU ablation for hysteromyoma [18,19].

Therefore, in this study, we aimed to develop and validate a combined model based on contrast-enhanced T1-weighted imaging (T1C) and T2 sequences for predicting the efficacy of HIFU ablation for hysteromyoma.

Materials and methods

Patients

This retrospective study was approved by the Institutional Review Board, and the requirement for informed consent was waived (IRB no. 2022ER101-1). From November 2019 to August 2021, 258 patients underwent hysteromyoma ablation at the HIFU center. The inclusion criteria were as follows: (I) premenopausal or perimenopausal women above 18 years old diagnosed with hysteromyoma by clinical and radiological examination; (II) patients undergoing their first HIFU ablation of hysteromyoma and having 10 or fewer hysteromyomas; (III) MRI examination within three days before and after ablation treatment; and (IV) no previous history of surgery or drug treatment. The exclusion criteria were as follows: (I) other uterine or accessory diseases; (II) lack of some MRI sequences or images; (III) poor MR image quality; and (IV) hysteromyomas with diameter ≤3 cm or diameter ≥10 cm [20]. According to the inclusion and exclusion criteria, 142 eligible patients (with 172 hysteromyomas; 45 (42–48) years) were enrolled in the study cohort. From January 2020 to July 2022, 40 patients (with 40 hysteromyomas; 43 (34–46) years) in another group were screened according to the same criteria. The patient recruitment process is shown in Figure 1.

Figure 1. Flowchart of patient recruitment with inclusion and exclusion criteria (a, b).

MRI examination

The MR images of the training and validation cohorts were from a GE Healthcare 3T MRI scanner, and the MR images of the independent test cohort were from a Unite Imaging Healthcare 3T MRI scanner. All patients underwent both T1C and T2 examinations before surgery. The MRI sequences and parameters are presented in Supplementary Table S1. The contrast agent was injected intravenously at a dose of 0.1 mmol/kg and a flow rate of 2.0 ml/s. Early, middle and late arterial images were collected approximately 15 s, 30 s and 45 s after the contrast agent was injected.

Patient grouping

The volume (V) and non-perfusion volume (NPV) (NPVR = NPV/V × 100%) of hysteromyomas were calculated on preoperative and postoperative T1C images using the ellipsoid formula (V = 0.5233 × D1 × D2 × D3 [21], where D1, D2 and D3 are the maximal longitudinal, anteroposterior and transverse diameters, respectively). Some studies have shown that the efficacy of HIFU ablation of hysteromyomas is related to the NPVR. The volume of hysteromyomas with an NPVR ≥70% is significantly reduced within one year after operation, and the cumulative clinical recurrence rate in two years is <10% (24, 25). Therefore, with NPVR = 70% as the boundary, hysteromyomas in the training and validation cohorts were divided into a sufficient ablation group (NPVR ≥70%, n = 95) and an insufficient ablation group (NPVR <70%, n = 77), and hysteromyomas in the test cohort were divided into a sufficient ablation group (NPVR ≥70%, n = 27) and an insufficient ablation group (NPVR <70%, n = 13).

Clinical–radiological features

Clinical–radiological features that might affect the NPVR of hysteromyomas were analyzed, including age, volume of hysteromyoma, type of hysteromyoma, T2 signal intensity of hysteromyoma (hypointensity: signal intensity equal to that of the skeletal muscle; isointensity: higher than that of the skeletal muscle but lower than that of the myometrium; hyperintensity: similar to or higher than that of the myometrium [22]), enhancement degree of hysteromyoma (mild: lower than that of the myometrium; moderate: similar to that of the myometrium; obvious: higher than that of the myometrium [22]), enhancement signal homogeneity of hysteromyoma and color Doppler flow imaging (CDFI) signal (small amount: Adler grade I, meaning that one or two short rods or punctate flow signals are observed; slightly abundant: Adler grade II or Adler grade III. Adler grade II implies that three to four punctate vessels are observed; in Adler grade III, more than four blood vessels, or the blood vessels are intertwined into a network are observed [23]).

Tumor segmentation and radiomics feature extraction

3D Slicer (version 4.10.2, https://www.slicer.org) was used for tumor region of interest (ROI) segmentation and feature extraction. In the case of unknown hysteromyoma information, one radiologist manually segmented the preoperative T1C (late arterial) sequence images, and another radiologist manually segmented the preoperative T2 sequence images and delineated the hysteromyoma area ROI layer by layer, generating a 3D volume of interest (VOI) for hysteromyomas by fusing the ROIs of each layer of images. To evaluate the repeatability of radiomics features, approximately 1/3 of the images from the T1C and T2 sequences of all the hysteromyomas were randomly selected one month later, and the hysteromyomas were again delineated layer by layer by above two radiologists, respectively. The intergroup correlation coefficient (ICC) was used to evaluate the consistency between the measures. An ICC >0.75 was considered to indicate good consistency of radiomics features. The original images were transformed by Laplacian Gaussian filtering and wavelet transform filtering, and radiomic features of T1C and T2 were extracted from VOI_T1C and VOI_T2, respectively. A total of 1223 radiomic features were extracted from both VOI_T1C and VOI_T2, including first-order features, shape features, texture features (grey level cooccurrence matrix (GLCM), grey level dependence matrix (GLDM), grey level run length matrix (GLRLM), grey level size zone matrix (GLSZM) and neighboring grey tone difference matrix (NGTDM)). A total of 2446 features were extracted from the T1C and T2 images.

Feature selection

The R software (version 4.1.1, https://www.r-project.org) and uAI Research Portal (version 730) were used for feature screening. For the clinical–radiological features, univariate and multivariate logistic regression (LR) analyses were performed to test the correlation between clinical–radiological risk factors for hysteromyoma and the NPVR, and the features with the strongest correlation were screened as independent clinical–radiological risk factors. For the radiomics features of T1C and T2 images, Z score normalization was used to eliminate the dimensional effects of different features. Then, the variance threshold method was used to remove the features with variance lower than 0.8. The Select K Best method was then used to remove the features that did not significantly differ between the two groups. Finally, the least absolute shrinkage and selection operator (LASSO) was used to select the features highly correlated with the NPVR as the optimal radiomic features.

Model construction

A LR algorithm was used to establish three models: independent clinical–radiological risk factors were used to construct the clinical–radiological model, the optimal T1C and T2 features were used to construct the radiomics model, and three groups of features were incorporated to construct the combined model. These models comprehensively reflected the factors that influence the NPVR from different perspectives and compared and analyzed the ability of different models to predict the efficacy of HIFU ablation for hysteromyomas.

Model performance evaluation

A fivefold cross validation procedure was adopted to validate the model. The procedure randomly divided all hysteromyomas into five subgroups, including a validation cohort and a training cohort with four subgroups. The experiment was repeated five times. The receiver operating characteristic (ROC) curve was drawn, and the area under the ROC curve (AUC), sensitivity, specificity, accuracy and F1 score were calculated to quantify the predictive performance of each model for HIFU ablation of hysteromyomas in the training cohort and validation cohort. Integrated discrimination improvement (IDI) was used to compare the improvement degree of the prediction accuracy of the different models. A nomogram was constructed using the combined model based on LR in the training cohort. The Hosmer–Lemeshow (H–L) test evaluated the consistency between the actual HIFU ablation efficacy and the predicted HIFU ablation efficacy and was used to draw the calibration curves of the training cohort and validation cohort. Decision curve analysis (DCA) was performed to evaluate the clinical usefulness of the model by quantifying the net benefits at different threshold probabilities (Figure 2). An independent test cohort was applied to test the performance of the optimal model.

Figure 2. Flowchart showing the process for the development of radiomics.

Statistical analysis

The R language package (R Foundation for Statistical Computing, Vienna, Austria) and IBM SPSS Statistics 26.0 (Armonk, NY) were used for statistical analysis. Quantitative data are expressed as ($\overline{\chi }$ ± ѕ) or M (Q₂₅ – Q₇₅) and were compared with the independent sample t-test or Wilcoxon’s rank sum test, and qualitative data are expressed as percentages and compared with the Chi-square test or Fisher’s exact test. A two-sided p value of 0.05 was used to determine statistical significance.

Results

Clinical–radiological features

The clinical–radiological features of the study patients are shown in Table 1. In the training and validation cohorts, univariate analysis indicated that T2 signal intensity, CDFI signal and enhancement degree were significantly different between the sufficient ablation group and insufficient ablation group. Multivariate analysis indicated that T2 signal intensity (odds ratio, OR = 1.064; p = 0.044) and CDFI signal (OR = 0.045; p = 0.007) were independent clinical–radiological risk factors for predicting the NPVR (Table 2).

Table 1. Clinical–radiological features of patients in the training, validation and independent test datasets.

Clinical–radiological features	Training and validation cohorts (n = 172)			Test cohort (n = 40)			p
	Insufficient ablation (n = 77)	Sufficient ablation (n = 95)	p	Insufficient ablation (n = 13)	Sufficient ablation (n = 27)	p
Age (years)^c	45 (42–49)	45 (41–47)	0.292a	43 (36–47)	36 (33–43)	0.189a	<0.001^a
Volume^c	37983.78 (23710.59–90801.70)	73943.71 (45071.00–130346.97)	<0.001^a	47890.66 (23388.17–108060.77)	53925.14 (25359.10–107848.31)	0.864^a	0.683^a
Type			0.770b			0.574^b	0.990^b
Subserous	13 (16.88)	8 (8.42)		0 (0.00)	2 (7.41)
Intramural	63 (81.82)	85 (89.47)		11 (84.62)	22 (81.48)
Submucous	1 (1.30)	2 (2.11)		2 (15.38)	3 (11.11)
T2WI signal intensity			<0.001^b			0.950^b	<0.001^b
Hyperintensity	34 (44.16)	18 (18.95)		6 (46.15)	3 (11.11)
Isointensity	13 (16.88)	10 (10.52)		4 (30.77)	12 (44.44)
Hypointensity	30 (38.96)	67 (70.53)		3 (23.08)	12 (44.44)
Enhancement degree			0.004^b			0.989^b	0.752^b
Mild	25 (32.47)	48 (50.53)		3 (23.08)	15 (55.56)
Moderate	27 (35.06)	35 (36.84)		8 (61.54)	4 (14.81)
Obvious	25 (32.47)	12 (12.63)		2 (15.38)	8 (29.63)
Enhancement signal homogeneity			0.766^b			1.000^b	0.622^b
Homogeneous	21 (27.27)	24 (25.26)		4 (30.77)	8 (29.63)
Inhomogeneous	56 (72.73)	71 (74.74)		9 (69.23)	19 (70.37)
CDFI signal			<0.001^b			<0.001^b	0.054^b
Small amount	60 (77.92)	93 (97.89)		5 (38.46)	26 (96.30)
Slightly abundant	17 (22.08)	2 (2.11)		8 (61.54)	1 (3.70)

Data are shown as number of patients and percentage in parentheses, unless otherwise stated.

^a p Values obtained using Wilcoxon’s rank-sum test.

^b p Values obtained using chi-squared test or Fisher’s exact test.

^c Data are medians and quartiles in parentheses.

Table 2. Univariate and multivariate analyses results of clinical–radiological features.

Clinical–radiological features	Univariate		Multivariate
	OR (95% CI)	p Value	OR (95% CI)	p Value
Age	0.453 (0.080–2.279)	0.348
Volume	3.742 (0.484–80.002)	0.283
Type	1.724 (0.923–3.348)	0.093
T2WI signal intensity	1.106 (1.047–1.171)	<0.001	1.064 (1.002–1.131)	0.044
Enhancement degree	0.810 (0.686–0.952)	0.012	0.851 (0.712–1.013)	0.072
Enhancement signal homogeneity	1.072 (0.676–1.697)	0.766
CDFI signal	0.024 (0.002–0.155)	0.001	0.045 (0.003–0.325)	0.007

Selection of radiomics features

In total, 2446 radiomics features were extracted for each patient. From these features, we selected 854 and 805 features with high consistency (both interobserver ICCs >0.75) for T1C and T2 images, respectively. After screening by the variance threshold method, selecting the best k method and the LASSO algorithm, 6 and 3 features remained in the T1C and T2 sequences, respectively (Supplementary Table S2).

Performance validation

The combined model yielded preferable predictive performance in both the training cohort (an average AUC = 0.841, 95% confidence interval (CI), 0.772–0.909) and validation cohort (an average AUC = 0.834, 95% CI, 0.691–0.971) compared with the clinical–radiological model (training cohort: an average AUC = 0.693, 95% CI, 0.611–0.775; validation cohort: an average AUC = 0.689, 95% CI, 0.525–0.853). The radiomics model also showed greater predictive performance than the clinical–radiological model in the training cohort (an average AUC = 0.803, 95% CI, 0.726–0.881), but not in the validation cohort (an average AUC = 0.794, 95% CI, 0.631–0.954).The IDI results showed that the combined model improved the prediction accuracy by 10.40% (training cohort, p < 0.001) and 14.1% (validation cohort, p < 0.001) compared with the clinical–radiological model, and it also improved the prediction accuracy by 2.9% (training cohort, p < 0.001) and 5.3% (validation cohort, p = 0.004) compared with the radiomics model (Figure 3, Tables 3 and 4, Supplementary Table S3).

Figure 3. ROC curves of clinical–radiological models, radiomics models and combined models in the training (a, c, e) and validation (b, d, f) cohorts.

Table 3. Performance of clinical–radiological, radiomics and combined models in the training and validation cohorts.

Model	AUC (95% CI)	Sensitivity	Specificity	Accuracy	Precision	F1 score
Clinical–radiological model	0.693 (0.611–0.775)|0.689 (0.525–0.853)	0.800|0.800	0.481|0.482	0.657|0.657	0.655|0.655	0.720|0.720
Radiomics model	0.803 (0.726–0.881)|0.794 (0.631–0.954)	0.782|0.758	0.705|0.676	0.747|0.721	0.765|0.744	0.773|0.751
Combined model	0.841 (0.772–0.909)|0.834 (0.691–0.971)	0.837|0.832	0.737|0.716	0.792|0.780	0.797|0.785	0.817|0.808

Training cohort|validation cohort.

Table 4. Comparison of models in the training and validation cohorts.

Model	Group	NRI (95% CI)	p (NRI)	IDI (95% CI)	p (IDI)
Combined model vs. clinical–radiological model	Training cohort	0.314 (0.144–0.484)	<0.001	0.104 (0.071–0.137)	<0.001
	Validation cohort	0.253 (–0.093 to 0.598)	0.152	0.141 (0.075–0.208)	<0.001
Combined model vs. radiomics model	Training cohort	0.098 (–0.016 to 0.212)	0.092	0.029 (0.012–0.045)	<0.001
	Validation cohort	0.119 (–0.098 to 0.337)	0.282	0.053 (0.017–0.089)	0.004

Net reclassification improvement (NRI) >0 and IDI >0 were positive improvement, indicating that the predictive ability of the new model was better than the old one.

Development and verification of the nomogram

The nomogram was constructed with optimal radiomics features and clinical–radiological variables as a convenient visible tool to estimate the risk of HIFU (Figure 4(A)). The calibration curves of the combined model demonstrated good consistency between the predictive outcome and observation in the training and validation cohorts (p = 0.251) (Figure 4(B,C)). The DCA plot indicated that the combined model had the highest clinical net benefit with wider threshold probabilities compared with the other models (Figure 4(D)). If the threshold probability is between 0.06 and 0.8, the application of the combined model to predict the successful clinical outcome of HIFU treatment provides more net benefits than the ‘treat-all patients’ or the ‘treat-none’ method.

Figure 4. The combined nomogram for predicting the efficacy of HIFU for hysteromyoma (a). The developed nomogram based on the combined model to predict the risk of insufficient HIFU ablation. T2WI signal intensity: 1, hyperintensity; 2, isointensity; 3, hypointensity; CDFI signal: 1, small amount; 2, slightly abundant. Calibration curves for the nomogram in the training and validation cohorts (b, c). The black dashed line represents the ideal prediction, and the green, red and purple lines represent the predictive ability of clinical–radiological models, radiomics models and combined models. The closer they fit to the dashed green line, the greater the prediction accuracy of the model. Decision curve analysis for the nomogram (d). The black line represents the net benefit of assuming no hysteromyoma patients receive HIFU ablation. The grey line is the net benefit of assuming that all hysteromyoma patients receive HIFU ablation. The green line, red line and purple line represent the expected net benefit of predicting stroke outcome using the clinical–radiological models, radiomics models and combined models, respectively.

Independent test

The proposed preoperative combined model was tested with independent data, and the model performed well, with an AUC of 0.801 (sensitivity: 81.5%; specificity: 61.5%; accuracy: 75.0%).

Discussion

In this study, we used the LR classifier to develop and validate three models: the clinical–radiological model, radiomics model and combined model. According to the results, the predictive performance of the combined model outperformed those of the radiomics model and clinical–radiological model. To our knowledge, this study is the first to predict the efficacy of HIFU ablation for hysteromyoma based on T1C combined with T2 radiomics, which will help clinicians judge the difficulty of ablation before surgery, select an appropriate treatment plan for patients, and promote the development of personalized medicine.

Predicting the efficacy of HIFU based on traditional imaging parameters remains challenging. In our study, CDFI signal and T2 signal intensity were independent clinical–radiological risk factors for predicting the efficacy of HIFU ablation. However, the prediction efficiency of the clinical–radiological model based only on CDFI and T2 signal intensity was low. Although CDFI can reflect the vascular distribution and blood flow velocity of hysteromyoma and T2 signal intensity can reflect the density of hysteromyoma cells [24], the qualitative features were greatly affected by the subjective factors of the observer in the evaluation, which increased the uncertainty of the prediction results. At the same time, these signals were also affected by tissue type, main magnetic field intensity and magnetic field uniformity.

Therefore, a quantitative and reliable model needs to be found to predict the ablation efficacy of HIFU. Our study found that the combined model showed greater predictive performance than the clinical–radiological model. We propose three reasons for this improvement. First, radiomics can reveal the heterogeneity of tumor spatial distribution, and texture features reflect the pathological changes of tissues to an extent [25]. The small tissue differences caused by the pathological changes of a hysteromyoma in the image signal, shape and texture can be quantified by radiomics to overcome the subjectivity of image interpretation. Some studies have shown that radiomics may become a biomarker based on artificial intelligence to noninvasively capture the heterogeneity of hysteromyomas [26]. Second, T2 sequencing is also an important basic sequence for the evaluation of hysteromyoma because its signal changes are related to the pathological classification of myomas [27,28]. Previous studies [18] have confirmed the value of radiomics models based on T2 sequences in predicting the efficacy of HIFU. However, the T2 sequence could not directly reflect the blood flow of hysteromyoma and could not distinguish between degenerative myoma and cellular myoma-both growths showed high signals in the T2 sequence [29]. Cellular myoma was resistant to HIFU, which made HIFU energy difficult to deposit, and the NPVR was consequently low. In contrast, the number and blood supply of degenerative myoma cells decreased, making them easier to ablate. In contrast-enhanced MRI (CE-MRI), the blood supply of cellular myoma was rich; therefore, the degree of enhancement was high, while the blood supply of degenerative fibroids was reduced, resulting in a low degree of enhancement. Due to the different textural features, cellular myoma and degenerative fibroids are easy to distinguish. According to Penne’s bioheat transfer equation, focused ultrasound energy is lost with blood flow, causing the target temperature to be lower than the ablation set temperature [30,31]. CE-MRI can directly reflect the perfusion and permeability of myoma vessels and has potential clinical value in predicting and evaluating the NPVR [32]. Therefore, we included T1C sequences in this study to improve the characterization of myoma by the model, which might also be one of the reasons for the high efficiency of the combined model. Previous studies have also confirmed that the combination of sequences could fully reflect the information of tumors [33,34].

Although the predictive performance did not significantly differ between the combined model and the radiomics model, the IDI results showed that the prediction accuracy of the combined model was significantly improved. The IDI value reflects the overall improvement of the model. The nomogram based on the combined model showed good performance, calibration and clinical utility, with potential clinical significance in predicting the efficacy of HIFU ablation for hysteromyoma.

Several limitations to our study should be noted. First, this study had a single-center retrospective design with a small sample size. Before clinical application, we need to analyze a larger external dataset for further validation. Second, the laboratory index of patients was not recorded and analyzed with regard to clinical–radiological features. Finally, this study only extracted the features of late arterial images, which may limit the type and number of features. In the future, we intend to develop a multisequence radiomics model based on multicenter imaging and more clinical information.

Conclusion

The combined model based on dual-sequence MRI radiomics has the potential to predict the efficacy of HIFU ablation for hysteromyoma. In the future, further developing and verifying the model will help clinicians better screen patients who will benefit the most from HIFU treatment, provide a reference for treatment decision-making and allow for the formulation of accurate treatment plans.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data used to support the findings of this study are available from the corresponding author upon request

Additional information

Funding

This study has received funding by Nanchong City School Cooperation Project [No. 19SXHZ0429].