The developers of the Internet-based survival-prediction tool say that it “serves as a counseling and decision aid to patients and …assists in risk modeling.”
They explain in their report in Cancer online November 5 that the response to neoadjuvant treatment can vary widely even in patients with a similar clinical pretreatment disease stage (cTNM), and identifying patients who will benefit from aggressive treatment versus those who will be exposed to needless morbidity “continues to be a challenge.”
The investigators therefore aimed to create a tool to predict individualized survival probability with or without CRT, based on a patient’s particular pathologic, demographic and treatment data.
For this purpose, Dr. Charles R. Thomas and colleagues at Oregon Health & Science University, Portland analyzed outcomes in 824 patients over 65 years of age in the Surveillance, Epidemiology, and End Results (SEER)-Medicare database, who were diagnosed with esophageal cancer between 1997 and 2005. While all underwent esophagectomy, 262 of the patients (32%) also received neoadjuvant CRT.
The covariates examined included sex, T and N classification, histology, total number of lymph nodes examined, and treatment with esophagectomy or CRT followed by esophagectomy. The data did not include preoperative staging information or final surgical margin status.
Using propensity score weighting and regression analysis, the team produced coefficients for survival with esophagectomy alone versus CRT plus esophagectomy for all of the values of the covariates, and entered them in an online nomogram.
The paper includes a representative example, for a 70-year-old man with T3N0 disease in whom 12 lymph nodes were examined: in this case, the predicted median survival was 19 months with surgery alone verus 35 months with CRT plus surgery, and the predicted 3-year overall survival was 27% versus 49% under the two approaches, respectively.
“Our online tool is currently available for use and can be found at skynet.ohsu.edu/nomograms,” Dr. Thomas and colleagues state.
Overall, they conclude, “Although the predictive tool described in the current study does suffer from retrospective and pathologic staging limitations, it provides a statistical, usable, and patient-friendly blue print for predicting survival based on treatment and patient-specific clinical variables.”