Dataset Information

Predicting pathologic complete response in locally advanced rectal cancer patients after neoadjuvant therapy: a machine learning model using XGBoost.

ABSTRACT:

Purpose

Watch and wait strategy is a safe and effective alternative to surgery in patients with locally advanced rectal cancer (LARC) who have achieved pathological complete response (pCR) after neoadjuvant therapy (NAT); present restaging methods do not meet clinical needs. This study aimed to construct a machine learning (ML) model to predict pCR preoperatively.

Methods

LARC patients who received NAT were included to generate an extreme gradient boosting-based ML model to predict pCR. The group was divided into a training set and a tuning set at a 7:3 ratio. The SHapley Additive exPlanations value was used to quantify feature importance. The ML model was compared with a nomogram model developed using independent risk factors identified by conventional multivariate logistic regression analysis.

Results

Compared with the nomogram model, our ML model improved the area under the receiver operating characteristics from 0.72 to 0.95, sensitivity from 43 to 82.2%, and specificity from 87.1 to 91.6% in the training set, the same trend applied to the tuning set. Neoadjuvant radiotherapy, preoperative carbohydrate antigen 125 (CA125), CA199, carcinoembryonic antigen level, and depth of tumor invasion were significant in predicting pCR in both models.

Conclusion

Our ML model is a potential alternative to the existing assessment tools to conduct triage treatment for patients and provides reference for clinicians in tailoring individual treatment: the watch and wait strategy is used to avoid surgical trauma in pCR patients, and non-pCR patients receive surgical treatment to avoid missing the optimal operation time window.

SUBMITTER: Chen X

PROVIDER: S-EPMC9262764 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:PurposePathologic complete response (pCR) is a critical factor in determining whether patients with rectal cancer (RC) should have surgery after neoadjuvant chemoradiotherapy (nCRT). Currently, a pathologist's histological analysis of surgical specimens is necessary for a reliable assessment of pCR. Machine learning (ML) algorithms have the potential to be a non-invasive way for identifying appropriate candidates for non-operative therapy. However, these ML models' interpretability remains challenging. We propose using explainable boosting machine (EBM) to predict the pCR of RC patients following nCRT.MethodsA total of 296 features were extracted, including clinical parameters (CPs), dose-volume histogram (DVH) parameters from gross tumor volume (GTV) and organs-at-risk, and radiomics (R) and dosiomics (D) features from GTV. R and D features were subcategorized into shape (S), first-order (L1), second-order (L2), and higher-order (L3) local texture features. Multi-view analysis was employed to determine the best set of input feature categories. Boruta was used to select all-relevant features for each input dataset. ML models were trained on 180 cases from our institution, with 37 cases from RTOG 0822 clinical trial serving as the independent dataset for model validation. The performance of EBM in predicting pCR on the test dataset was evaluated using ROC AUC and compared with that of three state-of-the-art black-box models: extreme gradient boosting (XGB), random forest (RF) and support vector machine (SVM). The predictions of all black-box models were interpreted using Shapley additive explanations.ResultsThe best input feature categories were CP+DVH+S+R_L1+R_L2 for all models, from which Boruta-selected features enabled the EBM, XGB, RF, and SVM models to attain the AUCs of 0.820, 0.828, 0.828, and 0.774, respectively. Although EBM did not achieve the best performance, it provided the best capability for identifying critical turning points in response scores at distinct feature values, revealing that the bladder with maximum dose >50 Gy, and the tumor with maximum2DDiameterColumn >80 mm, elongation <0.55, leastAxisLength >50 mm and lower variance of CT intensities were associated with unfavorable outcomes.ConclusionsEBM has the potential to enhance the physician's ability to evaluate an ML-based prediction of pCR and has implications for selecting patients for a "watchful waiting" strategy to RC therapy.

Project description:Background: MicroRNAs (miRNAs) are small, non-coding, RNA molecules which regulate numerous cellular processes. Specific miRNA may be abnormally down-regulated or up-regulated in colorectal cancer and have been found associated with prognosis or response to treatments. However, no study has ever addressed their predictive role in rectal cancer. Therefore, we used microarray technology and RT-PCR to profile miRNA expression patterns in patients (pts) with rectal cancer, with the aim to identify a specific M-bM-^@M-^\signatureM-bM-^@M-^] associated with pathological complete response after neoadjuvant chemo-radiotherapy. Methods: 38 pts with locally advanced rectal cancer (cT3-4/N+) were treated with capecitabine-oxaliplatin and pelvic conformal radiotherapy (45 cGy) followed by surgery (after 6-8 weeks). Pathologic response was scored according to the tumor regression grade (TRG) scale. MiRNA expression profile was analysed by microarray on fresh frozen biopsies obtained before treatment start and confirmed by RT-PCR. The correlation between miRNA expression profile and the TRG coded as TRG1 (pathologic Complete Response-pCR) versus TRG >1 (no pCR) was assessed by statistical analysis methods specifically designed for this study. Findings: 14 miRNAs were selected by arrays analysis as differentially expressed in TRG1 pts and 13 were confirmed by RT-PCR. In particular, 11 miRNAs (miR-1183, miR-483-5p, miR-622, miR-125a-3p, miR-1224-5p, miR-188-5p, miR-1471, miR-671-5p, miR-1909*, miR-630 and miR-765) were significantly up-regulated in TRG1 pts, while 2 miRNAs were under -expressed (miR-1274b and miR-720). miR-622 and miR-630 showed 100% sensitivity and specificity in selecting TRG1 cases and were significantly correlated with EGFR (M-OM-^G2=11M-bM-^@M-"793; p= 0M-bM-^@M-"001) and TS expression (M-OM-^G2=10M-bM-^@M-"589; p= 0M-bM-^@M-"001). Interpretation: A set of 13 miRNAs is strongly associated with pathologic complete response and may represent a specific marker of response to chemo-radiotherapy in locally advanced rectal cancer. Experiments were performed on purified RNA from preoperatory biopsies of 38 patients with histological diagnosis of rectal adenocarcinoma invading through the intestinal wall and/or with pelvic lymph node involvement as evaluated by endorectal ultrasonography (uT3-T4 and/or uN+). Patients were treated with neoadjuvant chemo-radiotherapy (capecitabine + oxaliplatin in combination with 45 Gy of pelvic conformal radiotherapy). Patients have been divided in the following two groups: group A those who had obtained a pathologic complete response M-bM-^@M-^S TRG 1, including the patient without detectable disease who refused surgery, group B any pathologic response other than complete.