Clinical T Category of Non–Small Cell Lung Cancers: Prognostic Performance of Unidimensional versus Bidimensional Measurements at CT

Purpose To compare the prognostic performances of clinical T categorization between the longest diameter and average diameter at CT in patients who underwent surgical resection of non-small cell lung cancers (NSCLCs). Materials and Methods This study retrospectively determined clinical T categories based on the longest diameter (clinical T longest ) and average diameter (clinical T average ) in 1153 patients, including 651 men (median age, 67 years; interquartile range [IQR], 60-72 years) and 502 women (median age, 63 years; IQR, 55-70 years) who underwent preoperative chest CT and subsequent resection of NSCLCs (clinical T1 to clinical T4; N0M0) between 2009 and 2015. Prognostic performances for disease-free survival (DFS) were compared between clinical T longest and clinical T average by using the Harrell concordance indexes and Student t test. The effect of the average diameter on clinical T category shifts (downstaging) was also investigated by using the McNemar-Bowker test. Results Concordance indexes did not significantly differ between clinical T longest (0.72; 95% confidence interval [CI]: 0.67, 0.76) and clinical T average (0.70; 95% CI: 0.64, 0.75) (P = .12 for the comparison). In the clinical T1 subgroup analysis, concordance indexes were 0.77 (95% CI: 0.71, 0.83) for clinical T longest and 0.75 (95% CI: 0.69, 0.81) for clinical T average (P = .17 for the comparison). Use of the clinical T average resulted in significant downstaging in all clinical T categories (P < .001). Conclusion The prognostic performance of clinical T categorization was not significantly different between the longest and average diameter measurements. Clinical T categorization based on the longest tumor diameter at CT was demonstrated to be sufficient for risk stratification of surgically treated non-small cell lung cancers. © RSNA, 2019 Online supplemental material is available for this article.