Table 4 Areas under the receiver operating characteristic curve (AUROCs) for NLP models that interpret medical oncologist progress note text to ascertain clinical outcomes, as evaluated in the labeled test set.

Clinical outcome	Models trained on all cancer types
Clinical outcome	All patients	Breast cancer	Colorectal cancer	NSCLC	Pancreatic cancer	Prostate cancer	Renal cell carcinoma	Urothelial carcinoma
Any cancer	0.93	0.98	0.98	0.95	0.78	0.91	0.91	0.95
Progression	0.92	0.97	0.91	0.96	0.92	0.87	0.86	0.78
Response	0.93	0.96	0.95	0.95	0.91	0.87	0.89	0.99
Models trained on all cancers except for the type under evaluation
Any cancer	†	0.98	0.97	0.95	0.79	0.85	0.90	0.95
Progression	†	0.96	0.90	0.95	0.91	0.84	0.84	0.78
Response	†	0.94	0.94	0.94	0.88	0.83	0.88	1.0

Quick links

Search