Prediction of false-negative extramural venous invasion in patients with rectal cancer using multiple mathematical models of diffusion-weighted imaging

A machine learning model based on preoperative multiparametric quantitative DWI can effectively predict the survival and recurrence risk of pancreatic ductal adenocarcinoma

PURPOSE

To develop a machine learning (ML) model combining preoperative multiparametric diffusion-weighted imaging (DWI) and clinical features to better predict overall survival (OS) and recurrence-free survival (RFS) following radical surgery for pancreatic ductal adenocarcinoma (PDAC).

MATERIALS AND METHODS

A retrospective analysis was conducted on 234 PDAC patients who underwent radical resection at two centers. Among 101 ML models tested for predicting postoperative OS and RFS, the best-performing model was identified based on comprehensive evaluation metrics, including C-index, Brier scores, AUC curves, clinical decision curves, and calibration curves. This model's risk stratification capability was further validated using Kaplan-Meier survival analysis.

RESULTS

The random survival forest model achieved the highest C-index (0.828/0.723 for OS and 0.781/0.747 for RFS in training/validation cohorts). Incorporating nine key factors-D value, T-stage, ADC-value, postoperative 7th day CA19-9 level, AJCC stage, tumor differentiation, type of operation, tumor location, and age-optimized the model's predictive accuracy. The model had integrated Brier score below 0.13 and C/D AUC values above 0.85 for both OS and RFS predictions. It also outperformed traditional models in predictive ability and clinical benefit, as shown by clinical decision curves. Calibration curves confirmed good predictive consistency. Using cut-off scores of 16.73/29.05 for OS/RFS, Kaplan-Meier analysis revealed significant prognostic differences between risk groups (p < 0.0001), highlighting the model's robust risk prediction and stratification capabilities.

CONCLUSION

The random survival forest model, combining DWI and clinical features, accurately predicts survival and recurrence risk after radical resection of PDAC and effectively stratifies risk to guide clinical treatment.

CRITICAL RELEVANCE STATEMENT

The construction of 101 ML models based on multiparametric quantitative DWI combined with clinical variables has enhanced the prediction performance for survival and recurrence risks in patients undergoing radical resection for PDAC.

KEY POINTS

This study first develops DWI-based radiological-clinical ML models predicting PDAC prognosis. Among 101 models, RFS is the best and outperforms other traditional models. Multiparametric DWI is the key prognostic predictor, with model interpretations through SurvSHAP.

Multiparametric quantitative diffusion weighted magnetic resonance imaging can effectively predict the response to neoadjuvant therapy in borderline resectable pancreatic ductal adenocarcinoma

PURPOSE

To investigate whether multiparametric quantitative diffusion weighted magnetic resonance imaging (DWI) can effectively predict the neoadjuvant therapy (NAT) response in borderline resectable pancreatic ductal adenocarcinoma (BRPC).

METHODS

The clinicopathological data, including tumor size, location, and CA19-9 values, as well as DWI parameters(ADC, D, and f values) from 72 patients with BRPC, were analyzed. The differences and changes in these factors before and after NAT were compared to identify those most accurately reflect the response to NAT. ROC analysis was used to evaluate the diagnostic efficacy, and Kaplan-Meier survival analysis explored the relationship between DWI parameters and prognosis. Subgroup survival analysis to further identify populations potentially benefiting from NAT based on multiparametric DWI.

RESULTS

After-NAT, the response group showed significantly higher ADC and D values and lower f values compared to the non-response group. The ΔADC (OR: 12.24, P = 0.013) emerged as the most important independent factor for tumor response, demonstrating the highest diagnostic accuracy for NAT response with an AUC of 0.936. Kaplan-Meier showed the high ADC value group, high D value group and low f value group were associated with better prognosis after NAT; and the ΔADC ≥ 0 group, ΔD ≥ 0 group, and Δf < 0 group was significantly associated with better prognosis. In addition, subgroup analysis suggested two groups of patients might potentially benefit from NAT.

CONCLUSIONS

Multiparametric quantitative DWI may offer valuable insights into the efficacy and prognosis of NAT in BRPC. These findings have the potential to support the evaluation and decision-making process for patients undergoing NAT.

Grading of clear cell renal cell carcinoma by using monoexponential, biexponential, and stretched exponential diffusion-weighted MR imaging

Objectives

To evaluate the diagnostic accuracy of monoexponential, biexponential and stretched-exponential diffusion-weighted imaging (DWI) models in the grading of clear cell renal cell carcinoma (ccRCC).

Materials and Methods

Fifty-one patients with pathologically proven ccRCC underwent DWI with fifteen b factors (0, 10, 30, 50, 70, 100, 150, 200, 300, 400, 600, 800, 1000, 1500, 2000 sec/mm²) on a 3.0T MR scanner. The isotropic apparent diffusion coefficient (ADC), true diffusion coefficient (ADCslow), pseudodiffusion coefficient (ADCfast), and fraction of perfusion (f) were derived from DWI using a biexponential model. The water diffusion heterogeneity index (α) and distributed diffusion coefficient (DDC) were derived from DWI using a stretched-exponential model. All values were calculated for the solid area of tumors and compared between high-grade and low-grade ccRCC. The Mann-Whitney U test and receiver operating characteristic (ROC) analysis were used for statistical analysis. The DeLong test was performed to compare the ROC curves.

Results

The mean ADC, DDC, ADCslow and α values were significantly lower in high-grade ccRCC than in low-grade ccRCC (P< 0.01). However, the ADCfast and f were not significantly different between the two groups (P > 0.05). According to the ROC analyses, the AUC for α was 0.941, which was significantly greater than those of the other parameters, with a sensitivity of 100% and a specificity of 84.2%. The DeLong test showed that there were significant differences in the ROCs among ADCfast/ADC, ADCfast/α, f/ADCslow, ADCfast/ADCslow, f/α, DDC/α, and f/ADC.

Conclusions

Diffusion-related parameters (ADC, DDC, ADCslow and α) could be used to distinguish between low- and high-grade ccRCC. The α derived from the stretched-exponential model may be the most promising parameter for grading ccRCC.

MRI Assessment of Extramural Venous Invasion Before and After Total Neoadjuvant Therapy for Locally Advanced Rectal Cancer and Its Association with Disease-Free and Overall Survival

BACKGROUND

Extramural venous invasion (EMVI) on baseline MRI is associated with poor prognosis in patients with locally advanced rectal cancer. This study investigated the association of persistent EMVI after total neoadjuvant therapy (TNT) (chemoradiotherapy and systemic chemotherapy) with survival.

METHODS

Baseline MRI, post-TNT MRI, and surgical pathology data from 175 patients with locally advanced rectal cancer who underwent TNT and total mesorectal excision between 2010 and 2017 were retrospectively analyzed for evidence of EMVI. Two radiologists assessed EMVI status with disagreement adjudicated by a third. Pathologic EMVI status was assessed per departmental standards. Cox regression models evaluated the associations between EMVI and disease-free and overall survival.

RESULTS

EMVI regression on both post-TNT MRI and surgical pathology was associated with disease-free survival (hazard ratio, 0.17; 95% confidence interval (CI), 0.04-0.64) and overall survival (hazard ratio, 0.11; 95% CI, 0.02-0.68). In an exploratory analysis of 35 patients with EMVI on baseline MRI, only six had EMVI on pathology compared with 18 on post-TNT MRI; these findings were not associated (p = 0.2). Longer disease-free survival was seen with regression on both modalities compared with remaining positive. Regression on pathology alone, independent of MRI EMVI status, was associated with similar improvements in survival.

CONCLUSIONS

Baseline EMVI is associated with poor prognosis even after TNT. EMVI regression on surgical pathology is common even with persistent EMVI on post-TNT MRI. EMVI regression on surgical pathology is associated with improved DFS, while the utility of post-TNT MRI EMVI persistence for decision-making and prognosis remains unclear.

Extramural vascular invasion nomogram before radical resection of rectal cancer based on magnetic resonance imaging

Purpose

This study verified the value of magnetic resonance imaging (MRI) to construct a nomogram to preoperatively predict extramural vascular invasion (EMVI) in rectal cancer using MRI characteristics.

Materials and methods

There were 55 rectal cancer patients with EMVI and 49 without EMVI in the internal training group. The external validation group consisted of 54 rectal cancer patients with EMVI and 55 without EMVI. High-resolution rectal T2WI, pelvic diffusion-weighted imaging (DWI) sequences, and dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) were used. We collected the following data: distance between the lower tumor margin and the anal margin, distance between the lower tumor margin and the anorectal ring, tumor proportion of intestinal wall, mrT stage, maximum tumor diameter, circumferential resection margin, superior rectal vein width, apparent diffusion coefficient (ADC), T2WI EMVI score, DWI and DCE-MRI EMVI scores, demographic information, and preoperative serum tumor marker data. Logistic regression analyses were used to identify independent risk factors of EMVI. A nomogram prediction model was constructed. Receiver operating characteristic curve analysis verified the predictive ability of the nomogram. P &lt; 0.05 was considered significant.

Result

Tumor proportion of intestinal wall, superior rectal vein width, T2WI EMVI score, and carbohydrate antigen 19-9 were significant independent predictors of EMVI in rectal cancer and were used to create the model. The areas under the receiver operating characteristic curve, sensitivities, and specificities of the nomogram were 0.746, 65.45%, and 83.67% for the internal training group, respectively, and 0.780, 77.1%, and 71.3% for the external validation group, respectively.

Data conclusion

A nomogram including MRI characteristics can predict EMVI in rectal cancer preoperatively and provides a valuable reference to formulate individualized treatment plans and predict prognosis.

Diffusion kurtosis imaging and MRI-detected extramural venous invasion in rectal cancer: correlation with clinicopathological prognostic factors

OBJECTIVE

To investigate the prognostic value of the diffusion kurtosis imaging (DKI)-derived parameters D value, K value, diffusion-weighted imaging (DWI) parameter apparent diffusion coefficient (ADC) value, and magnetic resonance imaging (MRI)-detected extramural venous invasion (EMVI) (mrEMVI) in rectal cancer patients.

METHODS

Forty patients who underwent MRI for rectal cancer were retrospectively evaluated. DKI-derived parameters D and K were measured using the Medical Imaging Interaction Toolkit. Conventional ADC values were measured from the corresponding DWI images. An experienced radiologist evaluated the mrEMVI status on MR images using the mrEMVI scoring system. An independent sample t-test or analysis of variance was used to analyze and compare the measurement data. The x² test or Fisher exact test was used for categorical variables. Receiver operating characteristic curves were used to assess the diagnostic performance of these parameters.

RESULTS

Among the 40 patients, MRI showed positive EMVI in 15 patients and negative EMVI in 25 patients. Positive mrEMVI status was associated with age, positive circumferential resection margin, pT-stage, lymphovascular invasion (LVI), distant metastasis, and serum carcinoembryonic antigen (CEA) level (P = 0.004-0.036). The dispersion coefficient (D) values and ADC values were significantly higher in the mucinous adenocarcinoma (MC) group than in the common adenocarcinoma (AC) group (P = 0.001), while kurtosis coefficient (K) values were lower in the MC group than in the AC group (P = 0.022). D values were significantly higher in the KRAS-mutated group than in the wild-type group (P < 0.05), whereas K values were lower in the KRAS-mutated group than in the wild-type group (P < 0.05). All three parameters (D, K, and ADC values) showed good diagnostic performance for discriminating MC from AC. Both the D and K values showed certain diagnostic performance for discriminating KRAS mutation.

CONCLUSION

DKI-derived parameters, conventional ADC values, and mrEMVI are associated with different histopathological prognostic factors. All DKI-derived parameters and conventional ADC values may distinguish MC from AC. DKI-derived parameters may also be used to discriminate KRAS mutation.

Whole-tumor amide proton transfer-weighted imaging histogram analysis to predict pathological extramural venous invasion in rectal adenocarcinoma: a preliminary study

OBJECTIVES

To evaluate amide proton transfer-weighted (APTw)-derived whole-tumor histogram analysis parameters in predicting pathological extramural venous invasion (pEMVI) positive status of rectal adenocarcinoma (RA).

METHODS

Preoperative MR including APTw imaging of 125 patients with RA (mean 61.4 ± 11.6 years) were retrospectively analyzed. Two radiologists reviewed each case's EMVI status based on the MR-based modified 5-point scale system with conventional MR images. The APTw histogram parameters of primary tumors were obtained automatically using whole-tumor volume histogram analysis. The independent risk factors markedly correlated with pEMVI-positive status were assessed using univariate and multivariate logistic regression analyses. Diagnosis performance was assessed by receiver operating characteristic curve (ROC) analysis. The AUCs were compared using the Delong method.

RESULTS

Univariate analysis demonstrated that MR-tumor (T) stage, MR-lymph node (N) stage, APTw-10%, APTw-90%, interquartile range, APTw-minimum, APTw-maximum, APTw-mean, APTw-median, entropy, kurtosis, mean absolute deviation (MAD), and robust MAD were significantly related to pEMVI-positive status (all p < 0.05). Multivariate analysis demonstrated that MR-T stage (OR = 4.864, p = 0.018), MR-N stage (OR = 4.967, p = 0.029), interquartile range (OR = 0.892, p = 0.037), APT-minimum (OR = 1.046, p = 0.031), entropy (OR = 11.604, p = 0.006), and kurtosis (OR = 1.505, p = 0.007) were the independent risk factors enabling prediction of pEMVI-positive status. The AUCs for diagnostic ability of conventional MRI assessment, the APTw histogram model, and the combined model (including APTw histogram and clinical variables) were 0.785, 0.853, and 0.918, respectively. The combined model outperformed the APTw histogram model (p = 0.013) and the conventional MRI assessment (p = 0.006).

CONCLUSIONS

Whole-tumor histogram analysis of APTw images combined with clinical factors showed better diagnosis efficiency in predicting EMVI involvement in RA.

KEY POINTS

• Rectal adenocarcinomas with pEMVI-positive status are typically associated with higher APTw-SI values. • APTw-minimum, interquartile range, entropy, kurtosis, MR-T stage, and MR-N stage are the independent risk factors for EMVI involvement. • The best prediction for EMVI involvement was obtained with a combined model of APTw histogram and clinical variables (area under the curve, 0.918).