|
Back to 2014 Annual Meeting Abstracts
Immunoscoring for Prognostic Assessment of Colon Cancer: a Novel Complement to Ultrastaging
Simon Lavotshkin*1, John R. Jalas2, Hitoe Torisu-Itakura1, Junko Ozao-Choy1, Rafay a. Haseeb1, Alexander Stojadinovic3, Zev Wainberg4, Anton Bilchik1
1The John Wayne Cancer Institute, Santa Monica, CA; 2Saint John's Health Center, Santa Monica, CA; 3Uniformed Services University of Health and Sciences, Bethesda, MD; 4UCLA School of Medicine, Santa Monica, CA
Introduction: Although AJCC/TNM staging remains the gold standard for prognostic assessment of colon cancer, it cannot explain variable outcomes among patients with the same stage of disease. Several groups have examined a prognostic immunoscore based on immune infiltrates in the primary tumor. We hypothesized that an immunoscore based on five immune variables might improve the accuracy of ultrastaging in patients with colon cancer.
Methods: Our study group comprised patients enrolled in an ongoing prospective trial of ultrastaging for colon cancer (RO1 CA090848). Resected tumor specimens were analyzed for CD3, CD4, CD8, CD68, and FoxP3 in a blinded fashion by a pathologist. Areas positive for tumor- infiltrating lymphocytes (TIL) were defined as hot spots and stratified as focal or diffuse based on their staining pattern on broad magnification. Hot spots were then scored as high or low, based on the briskness of the lymphocytic response, in the center of the tumor (CT) and in the invasive margin (IM). This categorical score was compared with a continuous score derived from analysis of 360 images on 36 patients with stage I-III colon cancer with ImageJ processing software. The immunoscore was then correlated with AJCC/TNM stage and with disease-free survival.
RESULTS: The mean number of nodes was 17. Fisher's exact test showed that the continuous variable scored by ImageJ analysis software matched the pathologist's categorical scoring system (p-value = 0.0048-0.0421 for all but CD68 and FoxP3). Mean TIL counts in the CT region were consistently higher in stage I than in stage III tumors: CD3, CD4 and CD8 counts were 774, 872 and 745, respectively, in stage I tumors, as compared with 535 (p=0.05), 614 and 487, respectively, in stage III tumors. This increase was not observed for CD68 or FoxP3. Patients with a disease-free survival >5 years tended to have a higher CD8/CD3 ratio in both IM and CT regions, as compared to patients with disease-free survival <36 months (p=0.08).
CONCLUSIONS: This is the first study to validate an immunoscore using specimens and data from a prospective clinical trial in which surgery and pathology techniques were standardized. Our preliminary results suggest that an immunoscore based on CD3, CD4 and CD8 corresponds with earlier stage colon cancer and improved disease-free survival and should be further examined for inclusion in the AJCC staging system.
Back to 2014 Annual Meeting Abstracts
|
