Survival prediction of colorectal liver metastases underwent surgical resection after neoadjuvant chemotherapy: Tumor response combined with the genetic and morphological evaluation score.

Introduction: Neoadjuvant chemotherapy is becoming routine for colorectal liver metastasis (CRLM) in patients with high risks of recurrence or in whom resection is difficult. This retrospective study aimed to establish a modified survival prediction model for patients with CRLM who underwent hepatectomy after neoadjuvant chemotherapy.

Materials and methods: A total of 619 patients who received neoadjuvant chemotherapy followed by hepatectomy between 2006 and 2021 were included and divided into training and validation groups at a ratio of 2:1. The model was established in training group and validated in validation group. Chemotherapy response was integrated into the genetic and morphological evaluation (GAME) score as a new NeoGAME model, with assigned points based on the hazard ratio in the multivariate Cox regression. The NeoGAME score grouping cutoff was divided using X-tile, and the predictive power was compared with that of traditional models.

Results: The 5-year overall survival were significantly different in the NeoGAME low-risk (0-2 points), medium-risk (3-4 points) and high-risk (≥5 points) groups (training group, P < 0.001; validation group, P = 0.0012). The area under the curve in predicting 5-year survival was 0.67 and 0.66 for the training and validation groups, respectively. Time-dependent receiver operating characteristic curve showed better discrimination ability of NeoGAME than the GAME score in predicting 5-year survival.

Conclusions: The newly established NeoGAME score can predict survival more precisely for patients with CRLM receiving neoadjuvant chemotherapy. Moreover, the model offers a useful tool for assessing tumor behavior and selecting a benefiting population for liver resection.