Cancer Medicine (Jun 2024)

A dynamic nomogram for predicting pathologic complete response to neoadjuvant chemotherapy in locally advanced rectal cancer

  • Guancong Wang,
  • Jiasen Li,
  • Ying Huang,
  • Yincong Guo

DOI
https://doi.org/10.1002/cam4.7251
Journal volume & issue
Vol. 13, no. 11
pp. n/a – n/a

Abstract

Read online

Abstract Aim To explore the clinical factors associated with pathologic complete response (pCR) for locally advanced rectal cancer (LARC) patients treated with neoadjuvant chemoradiotherapy (nCRT) and develop a web‐based dynamic nomogram. Methods Retrospective analysis of patients with examination confirmed LARC from 2011 to 2022. Patients from the Union Hospital of Fujian Medical University were included as the training cohort (n = 1579) and Zhangzhou Hospital of Fujian Medical University as the external validation cohort (n = 246). Results In the training cohort, after nCRT, 350 (22.2%) patients achieved pCR. More stomas were avoided in pCR patients (73.9% vs. 69.7%, p = 0.043). After a median follow‐up time of 47.7 months (IQR 2–145) shown OS (5‐year: 93.7% vs. 81.0%, HR = 0.310, 95%CI: 0.189–0.510, p < 0.001) and DFS (5‐year: 91.2% vs. 75.0%, HR = 0.204, 95%CI: 0.216–0.484, p < 0.001) were significantly better among patients with pCR than non‐pCR. Multivariable Logistic analysis shown pCR was significantly associated with Pre‐CRT CEA (HR = 0.944, 95%CI: 0.921–0.968; p < 0.001), histopathology (HR = 4.608, 95%CI: 2.625–8.089; p < 0.001), Pre‐CRT T stage (HR = 0.793, 95%CI: 0.634–0.993; p = 0.043), Pre‐CRT N stage (HR = 0.727, 95%CI: 0.606–0.873; p = 0.001), Pre‐CRT MRI EMVI (HR = 0.352, 95%CI: 0.262–0.473; p < 0.001), total neoadjuvant therapy (HR = 2.264, 95%CI: 1.280–4.004; p = 0.005). Meanwhile, the online version of the nomogram established in this study was publicized on an open‐access website (URL: https://pcrpredict.shinyapps.io/LARC2/). The model predicted accuracy with a C‐index of 0.73 (95% CI: 0.70–0.75), with an average C‐index of 0.73 for the internal cross validation and 0.78 (95% CI: 0.72–0.83) for the external validation cohort, showing excellent model accuracy. Delong test results showed the model has an important gain value for clinical characteristics to predict pCR in rectal cancer. Conclusions Patients with pCR had a better prognosis, including OS and DFS, and were independently associated with Pre‐CRT CEA, histopathology, Pre‐CRT T/N stage, Pre‐CRT MRI EMVI, and TNT. A web‐based dynamic nomogram was successfully established for clinical use at any time.

Keywords