Beyond Sedlis: A novel, histology-based nomogram for predicting recurrence risk and need for adjuvant radiation in cervical cancer—A NRG/GOG ancillary analysis.

Abstract

6019 Background: In GOG 49, factors associated with a 3-year, 30% recurrence risk in squamous cell carcinoma of the cervix (SCC) after surgery alone were defined. These "intermediate" risk factors [tumor size (TS), depth of tumor invasion (DOI), and lymphvascular space invasion (LVSI)] were then studied in GOG 92, which demonstrated the utility of treating patients (pts) with ≥2 intermediate risk factors with adjuvant radiation (RT), Sedlis Criteria. However, pts with < 30% recurrence risk were not eligible and few pts with adenocarcinoma (AC) were included. Our study purpose was 1) to evaluate recurrence risk factors for AC vs SCC, and 2) to define contemporary nomograms for adjuvant treatment in pts with both histologies. Methods: We performed an ancillary analysis of GOG 49, 92, and 141, and included Stage I pts who received no neoadjuvant/adjuvant therapy. Multivariable Cox proportional hazards models were created separately for AC and SCC to evaluate independent risk factors for recurrence. Model accuracy was tested with ROC curves. Prognostic nomograms were generated for 2-year recurrence risk for AC and SCC. Results: We identified 715 with SCC and 105 pts with AC; 142 with SCC (19.9%) and 18 with AC 17.1%) recurred. For SCC, factors associated independently with recurrence were: LVSI [HR 1.58 (CI 1.12-2.22)], DOI [middle 1/3, HR 4.31 (CI 1.81-10.26); deep 1/3, HR 7.05 (CI 2.99-16.64)] and TS [≥4cm HR 2.67 (CI 1.67-4.29)]. In contrast, for AC, only TS ≥4cm was independently associated with recurrence [HR 4.69 (CI 1.25-17.63)]. At 3 years, ROC curves for these models predicted recurrence with 76% and 75% accuracy for SCC and AC, respectively. Utilizing a nomogram generated from these models, for SCC, DOI had the greatest impact on recurrence, with mid 1/3 conferring an 18% risk and deep 1/3 a 32% risk, while LVSI and TS increased risk by 4-10%, respectively. In contrast, for AC, TS alone had the greatest impact on recurrence risk with TS 2-4cm conferring a 20% risk over 3 years and TS ≥4cm, a 28% risk. Conclusions: Our nomogram differs from the Sedlis Criteria in demonstrating that single, as well as a combination of risk factors predict substantial 3-year recurrence rates in Stage I cervical cancer. Furthermore, these factors differ by AC and SCC subtypes, suggesting that distinct, histology-specific nomograms may have greater utility in identifying pts who will most benefit from adjuvant therapy.