Predicting Upstage Rate As a Function of Delay from Diagnostic Imaging for Locally-Advanced Lung Cancer Radiotherapy

Abstract

This study combines data from three institutions to examine the rate of tumor progression as defined by comparing TNM (AJCC TNM, 7th edition) staging from diagnostic to that from treatment planning PET/CT images for patients with locally-advanced non-small cell lung cancer (LA-NSCLC). In this retrospective study of prospectively acquired data, TNM and overall stage records from patients treated for LA-NSCLC treated between 2004 and 2011were pooled from three international institutions. While data from 100 patients were pooled, only 78 provided full TNM staging. For each patient, changes in T, N, M (N=78), and overall stage (N=100) were recorded, as well as the time elapsed between the diagnostic and treatment planning PET/CT images. The inverse Kaplan-Meier product-limit method was used to define the probability of disease progression as a function of the time lapsed. The median time lapse between PET/CT scans for all 100 patients was 29 days (range: 2 – 341). Using the inverse Kaplan-Meier method, the upstage rate for a time lapse of 21 days (3 weeks) in T, N, M (N=78), and overall (N=100) stage were 2% (95% CI: 0-18%), 6% (2-15%), 7% (3-17%), and 8% (4-16)%, respectively. The converse rates for a 28-day (4 weeks) lapse were 4% (1-15%), 10% (4-21%), 8% (4-19%), and 13% (8-23%). The 95% CI reported here are smaller than those calculated from the individual cohorts. Our model demonstrates that the upstage rate for LA-NSCLC patients, over one month, is noteworthy. It is worrisome that, in the span of four weeks, upstage in M occurred for 8% of patients who are no longer eligible radical treatment and lowered expectation of survival. Therefore, LA-NSCLC patients require a reduction of the time elapsed between diagnostic imaging and radiotherapy; barring that, a treatment planning PET/CT scan should be acquired to ensure radiotherapy is planned using current data.