Table 3.

Concordance according to patient profile per LLM

Overall concordance per patient profile per LLM
Patient profiles		ChatGPT versions			Other LLM
		GPT3.5 Sept 21	GPT3.5 Jan 22	GPT4	Llama2	Bard
Postmenopausal luminal A N−	1	No	No	Yes	No	No
Postmenopausal luminal A N+	2	No	No	No	No	No
Premenopausal luminal A N−	3	Yes	Yes	Yes	No	No
Premenopausal luminal A N+	4	Yes	No	No	No	No
Postmenopausal luminal B Her2− N−	5	Yes	No	No	No	No
Postmenopausal luminal B Her2− N+	6	No	No	Yes	No	Yes
Premenopausal luminal B Her2− N−	7	No	No	No	No	Yes
Premenopausal luminal B Her2+N+	8	Yes	Yes	Yes	Yes	No
Postmenopausal Her2+ER/PR- N−	9	No	No	Yes	No	Yes
Postmenopausal Her2+ER/PR- N+	10	No	No	No	No	Yes
Premenopausal Her2+ER/PR- N−	11	Yes	Yes	Yes	Yes	No
Premenopausal Her2+ER/PR- N+	12	Yes	Yes	Yes	Yes	No
Postmenopausal triple negative N−	13	Yes	Yes	Yes	Yes	No
Postmenopausal triple negative N+	14	Yes	Yes	Yes	No	No
Premenopausal triple negative N−	15	Yes	Yes	Yes	Yes	No
Premenopausal triple negative N+	16	No	No	Yes	Yes	No
Postmenopausal DCIS, clear resection margin	17	No	No	No	No	No
Premenopausal DCIS, clear resection margin	18	No	No	No	No	No
Postmenopausal DCIS, narrow resection margin	19	No	No	No	No	No
Inflammatory breast cancer	20	Yes	No	Yes	No	No
		50.0%	35.0%	60.0%	30.0%	20.0%

LLM large language model, PP patient profile, N+ nodal positive, N− nodal negative, Her2+ Her2 positive, Her2− Her2 negative, DCIS ductal carcinoma in situ, Sept September, Jan January