American Surgical Association

Back to 2011 Program


Identification Of A Biomarker Profile Associated With Resistance To Neoadjuvant Chemoradiation Therapy In Rectal Cancer
Julio Garcia-Aguilar, MD, PhD*1, Zhenbin Chen, MD*1, David D Smith, MD*1, Wenyan Li, MD*1, Robert D Madoff, MD2
1City of Hope, Duarte, CA;2University of Minnesota, Minneapolis, MN

Objective: Some rectal cancer patients treated with neoadjuvant chemoradiation and surgery have no viable cancer cells in their surgical specimens achieving a pathologic complete response (pCR). These patients could potentially be spared an unnecessary operation. Identifying biomarkers of response will help select patients more likely to benefit from a non-surgical approach.
Methods: Patients with Ultrasound or MRI-staged II or III rectal cancer were treated with 5-FU and 50.4Gy radiation in a multicenter trial. pCR was defined according to AJCC criteria. DNA from pre-treatment tumor biopsies and post-treatment surgical specimens was screened for mutations (p53, K-ras, B-raf, CTNNB1, Mitochondria D310, PIK3A, APC) and polymorphisms (p53, CCND1, EGFR, TS, ERCC1, XDP, IL6, DPD, TLR2, VEGF) by PCR analysis and direct sequencing. Association of biomarkers individually or in combination with pCR was performed by Fisher’s Exact test and internally validated performing 100 iterations randomly selecting 70% of patients as training set and 30% as testing set.
Results: Of 132 patients, 33 (25%) had a pCR. K-ras mutation alone (p=0.01) or combined with p53 mutation was associated with lack of complete response. CCND1 gene polymorphism alone (p=0.04) or combined with both K-ras and p53 mutations was also associated with no pCR (Table 1).
Conclusions: Combination of gene mutations and polymorphisms are associated with resistance to CRT and could be used to select optimal surgical therapy in rectal cancer patients.
Table 1: Cross-validation and assay characteristics
p53; K-rasp53; K-ras; CCND1 AA
Prevalence27/132 (20%)43/132 (33%)
pCR03%
Fisher’s Exact p-value< 0.0001< 0.0001
ROC curve AUC0.630.70
Validity75% ± 0.6%76% ± 0.6%
Sensitivity %33/33 (100%)32/33 (97%)
Specificity %27/99 (27%)42/99 (42%)
Positive Predictive Value %33/105 (31%)32/89 (36%)
Negative Predictive Value %27/27 (100%)42/43 (98%)


Back to 2011 Program