Achieving a Cure Without Total Mesorectal Excision in Rectal Adenocarcinoma

ResNet-Vision Transformer based MRI-endoscopy fusion model for predicting treatment response to neoadjuvant chemoradiotherapy in locally advanced rectal cancer: A multicenter study

Background:

Neoadjuvant chemoradiotherapy followed by radical surgery has been a common practice for patients with locally advanced rectal cancer, but the response rate varies among patients. This study aimed to develop a ResNet-Vision Transformer based magnetic resonance imaging (MRI)-endoscopy fusion model to precisely predict treatment response and provide personalized treatment.

Methods:

In this multicenter study, 366 eligible patients who had undergone neoadjuvant chemoradiotherapy followed by radical surgery at eight Chinese tertiary hospitals between January 2017 and June 2024 were recruited, with 2928 pretreatment colonic endoscopic images and 366 pelvic MRI images. An MRI-endoscopy fusion model was constructed based on the ResNet backbone and Transformer network using pretreatment MRI and endoscopic images. Treatment response was defined as good response or non-good response based on the tumor regression grade. The Delong test and the Hanley–McNeil test were utilized to compare prediction performance among different models and different subgroups, respectively. The predictive performance of the MRI-endoscopy fusion model was comprehensively validated in the test sets and was further compared to that of the single-modal MRI model and single-modal endoscopy model.

Results:

The MRI-endoscopy fusion model demonstrated favorable prediction performance. In the internal validation set, the area under the curve (AUC) and accuracy were 0.852 (95% confidence interval [CI]: 0.744–0.940) and 0.737 (95% CI: 0.712–0.844), respectively. Moreover, the AUC and accuracy reached 0.769 (95% CI: 0.678–0.861) and 0.729 (95% CI: 0.628–0.821), respectively, in the external test set. In addition, the MRI-endoscopy fusion model outperformed the single-modal MRI model (AUC: 0.692 [95% CI: 0.609–0.783], accuracy: 0.659 [95% CI: 0.565–0.775]) and the single-modal endoscopy model (AUC: 0.720 [95% CI: 0.617–0.823], accuracy: 0.713 [95% CI: 0.612–0.809]) in the external test set.

Conclusion:

The MRI-endoscopy fusion model based on ResNet-Vision Transformer achieved favorable performance in predicting treatment response to neoadjuvant chemoradiotherapy and holds tremendous potential for enabling personalized treatment regimens for locally advanced rectal cancer patients.

Total neoadjuvant treatment using short-course radiotherapy and four CAPOX cycles in locally advanced rectal cancer with high-risk criteria for recurrence: a Swedish nationwide cohort study (LARCT-US)

Background

Total neoadjuvant treatment (TNT) for locally advanced rectal cancer (LARC) increases pathologic complete response (pCR) rate and reduces the risk of systemic recurrences over chemoradiotherapy (CRT) in randomised trials, e.g., the RAPIDO trial. A modified RAPIDO schedule was prospectively explored in Sweden to evaluate TNT in routine health care before the RAPIDO results were published.

Methods

Between July 2016 and June 2020, 273 patients with high-risk LARC (clinical tumour stage cT4, clinical nodal stage cN2, extramural vascular invasion, involved mesorectal fascia or enlarged lateral lymph nodes) were treated in a prospective observational cohort study at 16 hospitals (LARCT-US). Another 189 patients at 18 (including the 16) hospitals were similarly treated (ad modum LARCT-US, AdmL) during the same period. Inclusion and exclusion criteria were identical to the RAPIDO trial. Patients received short-course radiotherapy (5 × 5 Gy for 5 days) followed by four cycles of CAPOX or six FOLFOX-6, followed by total mesorectal excision or, if clinical complete response (cCR), inclusion into a watch-and-wait (W&W) study. The primary endpoint was complete response (CR), i.e., the sum of pCR in specimens and cCR exceeding one year in W&W patients. Safety was assessed in all patients.

Findings

Compared to the RAPIDO trial, patients were older, and tumours more advanced. Median follow-up was 4.8 years (IQR 4.2-5.2). In LARCT-US all patients received radiotherapy and 268 (98%) started chemotherapy whereas in AdmL all patients received radiotherapy and chemotherapy. In LARCT-US 34 patients had pCR and 31 sustained cCR resulting in a CR-rate of 24% (95% CI 20-28). In AdmL, results were similar (23%, 95% CI 17-30). Locoregional recurrences were 6% (95% CI 4-10) and 5% (95% CI 2-9), respectively, both at 3 years and at last follow-up. Neurotoxicity, recorded in LARCT-US, was lower than in RAPIDO (EORTC-QLQ-CIPN20 tingling toes or feet mean score 24 (SD 31) vs 43 (SD 37)). One treatment-associated death occurred.

Interpretation

Despite older patients and more advanced tumours, results similar to the RAPIDO trial were obtained. Hence, two chemotherapy cycles less do not compromise the results maintaining a high CR-rate. This TNT schedule resulted in favourable outcomes in a nation-wide real-life situation.

Funding

Swedish Cancer Society.

Arterial Mucosal Linear Enhancement at Contrast-enhanced MRI to Exclude Residual Tumor after Neoadjuvant Chemotherapy and Radiation Therapy for Rectal Cancer

Background A watch-and-wait regimen for locally advanced rectal cancer after neoadjuvant chemotherapy and radiation therapy (NCRT) relies on identifying complete tumor response. However, the concordance between a complete response at combined T2-weighted and diffusion-weighted MRI (T2DWI) and pathologic complete response (pCR; ie, ypT0N0) in the tumor is unsatisfactory. Purpose To assess whether identification of mucosal linear enhancement (MLE) at arterial-phase contrast-enhanced (CE) T1-weighted MRI is associated with ypT0 status in patients with locally advanced rectal cancer after NCRT and to evaluate whether combining MLE at CE T1-weighted MRI and negative lymph node metastasis (LNM) at T2DWI can improve identification of pCR. Materials and Methods This retrospective study included patients with locally advanced rectal cancer who underwent total mesorectal excision after NCRT between July 2020 and July 2023 at a tertiary referral academic center. Restaging MRI included T2DWI and arterial-phase CE T1-weighted MRI for primary tumor assessment and T2DWI for evaluation of LNM status. Imaging features associated with ypT0 status were identified at multivariable regression analysis. Results In total, 239 patients (mean age, 58 years ± 12 [SD]; 180 male patients) were assessed. MLE was more common in the ypT0 group than in the ypT1-4 group after NCRT (73% vs 4%, respectively; P < .001). MLE was associated with higher odds of ypT0 status in an adjusted analysis (odds ratio, 137; 95% CI: 25, 767; P < .001). The combination of MLE and negative LNM status achieved an area under the receiver operating characteristic curve of 0.84 (95% CI: 0.79, 0.88) for pCR. Conclusion MLE at CE MRI was associated with higher odds of complete tumor response. Combining MLE and negative LNM status showed good performance for identifying complete tumor response and may exclude residual tumors after NCRT in patients with locally advanced rectal cancer. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Schoellnast in this issue.

Deep learning model based on endoscopic images predicting treatment response in locally advanced rectal cancer undergo neoadjuvant chemoradiotherapy: a multicenter study

PURPOSE

Neoadjuvant chemoradiotherapy has been the standard practice for patients with locally advanced rectal cancer. However, the treatment response varies greatly among individuals, how to select the optimal candidates for neoadjuvant chemoradiotherapy is crucial. This study aimed to develop an endoscopic image-based deep learning model for predicting the response to neoadjuvant chemoradiotherapy in locally advanced rectal cancer.

METHODS

In this multicenter observational study, pre-treatment endoscopic images of patients from two Chinese medical centers were retrospectively obtained and a deep learning-based tumor regression model was constructed. Treatment response was evaluated based on the tumor regression grade and was defined as good response and non-good response. The prediction performance of the deep learning model was evaluated in the internal and external test sets. The main outcome was the accuracy of the treatment prediction model, measured by the AUC and accuracy.

RESULTS

This deep learning model achieved favorable prediction performance. In the internal test set, the AUC and accuracy were 0.867 (95% CI: 0.847-0.941) and 0.836 (95% CI: 0.818-0.896), respectively. The prediction performance was fully validated in the external test set, and the model had an AUC of 0.758 (95% CI: 0.724-0.834) and an accuracy of 0.807 (95% CI: 0.774-0.843).

CONCLUSION

The deep learning model based on endoscopic images demonstrated exceptional predictive power for neoadjuvant treatment response, highlighting its potential for guiding personalized therapy.

Neoadjuvant chemotherapy is noninferior to chemoradiotherapy for early-onset locally advanced rectal cancer in the FOWARC trial

BACKGROUND

The early-onset rectal cancer with rapidly increasing incidence is considered to have distinct clinicopathological and molecular profiles with high-risk features. This leads to challenges in developing specific treatment strategies for early-onset rectal cancer patients and questions of whether early-onset locally advanced rectal cancer (LARC) needs aggressive neoadjuvant treatment.

METHODS

In this post hoc analysis of FOWARC trial, we investigated the role of preoperative radiation in early-onset LARC by comparing the clinicopathological profiles and short-term and long-term outcomes between the early-onset and late-onset LARCs.

RESULTS

We revealed an inter-tumor heterogeneity of clinical profiles and treatment outcomes between the early-onset and late-onset LARCs. The high-risk features were more prevalent in early-onset LARC. The neoadjuvant radiation brought less benefits of tumor response and more risk of complications in early-onset group (pCR: OR = 3.75, 95% CI = 1.37-10.27; complications: HR = 11.35, 95% CI = 1.46-88.31) compared with late-onset group (pCR: OR = 5.33, 95% CI = 1.83-15.58; complications: HR = 5.80, 95% CI = 2.32-14.49). Furthermore, the addition of radiation to neoadjuvant chemotherapy didn't improve long-term OS (HR = 1.37, 95% CI = 0.49-3.87) and DFS (HR = 1.05, 95% CI = 0.58-1.90) for early-onset patients.

CONCLUSION

Preoperative radiation plus chemotherapy may not be superior to the chemotherapy alone in the early-onset LARC. Our findings provide insight into the treatment of early-onset LARC by interrogating the aggressive treatment and alternative regimens.

Recent advances in rectal cancer treatment - are we on the right track?

Background

Staging and treatment of rectal cancer have evolved over several decades with considerably fewer locoregional recurrences but no marked improved survival since systemic recurrence risks remain virtually unchanged. This development will briefly be summarised followed by a thorough discussion of two recent developments.

Methods

A systematic approach towards the literature is aimed at focusing on organ preservation and the delivery of all non-surgical treatments prior to surgery or total neoadjuvant treatment (TNT).

Results

Organ preservation, that is to defer surgery if the tumour happens to disappear completely after any pre-treatment given to locally advanced tumours to decrease recurrence risks has increased in popularity and is, if not universally, widely accepted. To give neo-adjuvant treatment to intentionally obtain a clinically complete remission to avoid surgery is practised in some environments but is mostly still experimental. TNT, that is to provide both radiotherapy and chemotherapy aimed at killing microscopic disease in the pelvis or elsewhere has been subject to several trials. Collectively, they show that the chance of achieving a complete response, pathologically or clinically, has approximately doubled, increasing the chance for organ preservation, and the risk of distant metastasis has decreased at least in some trials. The best schedule remains to be established.

Conclusions

To obtain substantial progress and also improve survival, the systemic treatments need to be improved even if preoperative delivery is more effective and better tolerated than postoperative. The locoregional treatment may be further optimised through better risk prediction.

Prospective Correlation of Magnetic Resonance Tumor Regression Grade With Pathologic Outcomes in Total Neoadjuvant Therapy for Rectal Adenocarcinoma

PURPOSE

Total neoadjuvant therapy (TNT) is a newly established standard treatment for rectal adenocarcinoma. Current methods to communicate magnitudes of regression during TNT are subjective and imprecise. Magnetic resonance tumor regression grade (MR-TRG) is an existing, but rarely used, regression grading system. Prospective validation of MR-TRG correlation with pathologic response in patients undergoing TNT is lacking. Utility of adding diffusion-weighted imaging to MR-TRG is also unknown.

METHODS

We conducted a multi-institutional prospective imaging substudy within NRG-GI002 (ClinicalTrials.gov identifier: NCT02921256) examining the ability of MR-based imaging to predict pathologic complete response (pCR) and correlate MR-TRG with the pathologic neoadjuvant response score (NAR). Serial MRIs were needed from 110 patients. Three radiologists independently, then collectively, reviewed each MRI for complete response (mriCR), which was tested for positive predictive value (PPV), negative predictive value (NPV), sensitivity, and specificity with pCR. MR-TRG was examined for association with the pathologic NAR score. All team members were blinded to pathologic data.

RESULTS

A total of 121 patients from 71 institutions met criteria: 28% were female (n = 34), 84% White (n = 101), and median age was 55 (24-78 years). Kappa scores for T- and N-stage after TNT were 0.38 and 0.88, reflecting fair agreement and near-perfect agreement, respectively. Calling an mriCR resulted in a kappa score of 0.82 after chemotherapy and 0.56 after TNT reflected near-perfect agreement and moderate agreement, respectively. MR-TRG scores were associated with pCR (P < .01) and NAR (P < .0001), PPV for pCR was 40% (95% CI, 26 to 53), and NPV was 84% (95% CI, 75 to 94).

CONCLUSION

MRI alone is a poor tool to distinguish pCR in rectal adenocarcinoma undergoing TNT. However, the MR-TRG score presents a now validated method, correlated with pathologic NAR, which can objectively measure regression magnitude during TNT.