TIL Density and Tumor Budding Validated as Prognostic Factors for DFS in Stage III CRC

Tumor-infiltrating lymphocyte (TIL) density and tumor budding have been validated as a prognostic variable and provide stratification by T and N risk groups to predict disease-free survival (DFS) in patients with stage III colorectal cancer (CRC) in the pivotal NCCTG N0147 study, according to HalioDx.¹

The results of the NCCTG N0147 study, which were published in the Annals of Oncology, strengthen the clinical value of the Immunoscore as a diagnostic tool that can enable personalized management of patients with stage III CRC.

“[The study confirmed] that use of Immunoscore® in combination with established prognostic factors allows one to more accurately assess prognosis, highlighting the clinical utility of Immunoscore®,” said Frank A. Sinicrope, MD, principal investigator of the study and fellow of Medical Oncology at the Mayo Clinic, in response to the published data.

The NCCTG N0147 study enrolled patients with surgically resected stage III colon adenocarcinomas (n = 1532) to receive adjuvant FOLFOX (5-fluorouracil, leucovorin, and oxaliplatin)-based chemotherapy with or without cetuximab (Erbitux). The median follow-up time on study was 83 months. Patients were randomly and equally divided into a training set (n = 766) and a validation set (n = 766). Patients who had low TIL densities compared with those who had high TIL densities had significantly poorer DFS in the training set (HR, 2.01; 95% CI, 1.41%-2.85%; P <.0001), which was confirmed in the validation set (HR, 1.53; 95% CI, 1.06%-2.21%; P -.0196).²

High tumor budding was significantly associated with poorer DFS in the training set as well versus high tumor budding (HR, 1.56; 95% CI, 1.19%-2.05%; P =.0002), confirmed in the validation set (HR, 1.24; 95% CI .094-1.64; P =.0307). The presence of tumor micropapillary features was prognostic in the training set compared with the absence of these features (HR, 1.42; 95% CI, 1.08%-1.91%), but this was not confirmed in the validation set (HR, 0.94; 95% CI, 0.68%-1.31%; P =.7232).

Out of 1,532 tumors, 1,123 had low TILs defined as ≤3/high power field (HPF; 74%), and 400 had high TILs defined as >3/HPF (26%). Patients with low TILs were significantly more likely to have a higher T stage compared with patients with high TILs (P=.0003), as well as N stage (P =.0001). Patients with low TILs also were significantly more likely to have poor or undifferentiated histologic grade (P =.0001), left-sided primary location (P <.0001), non-mutated BRAF V600E (P <.0001), and pMMR (P <.0001).

Overall, 699 tumors had low budding (46%), and 833 tumors had high budding (54%). Patients with high budding were significantly more likely to have a higher T stage (P =.0002) and N stage (P <.0001) compared with patients with low budding.

Micropapillary features were observed in 326 tumors (21%), and these patients were more likely to have a higher T stage (P =.0329) and N stage (P =.0010), as well as pMMR (P <.0001) compared with tumors that did not have micropapillary features.

Overall, 10.1% of tumors had dMMR, and low TIL density tumors were found in 37% of those with dMMR compared with 78% with pMMR. High budding was also observed in 37% of those with dMMR compared with 57% with pMMR. Micropapillary features were present in 14% of the tumors with dMMR compared with 22% with pMMR. Investigators noted that the KRAS mutational status and treatment arm were not associated with TIL density, tumor budding, or micropapillary features. BRAF, however, was associated with high TIL density, which was described by enrichment of BRAF V600E in dMMR tumors, but BRAF V600E was not associated with tumor budding or micropapillary features.

Patients with low TIL density compared with high TIL density had significantly poorer DFS (HR, 1.74; 95% CI, 1.35%-2.24%; P <.0001), and these tumors were also shown to prognostically stratify patients with dMMR (HR, 2.14; 95% CI, 1.14%-4.01%; P =.0173) and pMMR (HR, 1.65; 95% CI, 1.25%-2.16%; P =.0002). Patients with high tumor budding compared with low tumor budding had significantly poorer DFS (HR, 1.37; 95% CI, 1.13%-1.67%; P =.0011). The presence of micropapillary features compared with the absence of these features was not significantly associated with DFS (HR, 1.17; 95% CI, 0.94%-1.45%; P =.1548).

Investigators categorized patients to 3 subgroups, which included low budding/high TILs, intermediate, and high budding/low TILs to determine the association of their combined variable with DFS. The combined value was shown to significantly stratify patients for prognosis with 5-year DFS rates of 78% in the low budding/high TILs group, 70% in the intermediate group, and 60% for the high budding/low TILs group (P <.0001). These data compared with the individual variables appeared more robust for the prediction of DFS.

Tumor budding/TILs appeared to better stratify high-risk patients for DFS compared with the low-risk group. The 5-year DFS rates in the low-risk patients were 80% for low budding/high TILs, 80% for intermediate, and 71% for high budding/low TILs (P =.0273), and DFS rates among high-risk patients were 75%, 58%, and 51%, respectively (P<.0001). Patients with high budding/low TILs had the poorest DFS among both the low- and high-risk patients. The 5-year DFS rates were also lower in low-risk patients with high budding/low TILs (71%) compared with high-risk patients with low budding/high TILs (75%).

According to the analysis of the relative contribution of each tumor variable to DFS, the top 3 contributors in the multivariable model with TILs were N stage (44.3%), TILs (14.8%), and KRAS mutations (13.8%). In the multivariable model with tumor budding, the top contributors were N stage (47.6%), T stage (14.3%), KRAS (13.6%), and tumor budding (8.3%).

Overall, low TILs, high budding, and the presence of micropapillary features were significantly higher with T stage, N stage, and pMMR. TIL density and tumor budding also appeared to be significantly associated with DFS after adjustments for covariates. The prognostic significance was confirmed in the validation set of the study.

The study validated TILs and tumor budding as significant prognostic variables for patients with stage III CRC, and the combination of these factors can provide more precise stratification for prognosis among these groups of patients, Sinicrope et al concluded.

_Reference

1. _{New evidence from 2 pivotal studies reinforces clinical value of Immunoscore in Colon Cancer. News release. HalioDx. May 26, 2020. Accessed May 28, 2020. https://bit.ly/2X7Zdlw.}
2. _{Lee H, Sha D, Foster NR, et al. Analysis of tumor microenvironmental features to refine prognosis by T, N risk group in patients with stage III colon cancer (NCCTG N0147) (Alliance). Annals of Oncol. 2020. 31;4, 487-494. doi:10.1016/j.annonc.2020.01.011}