Breast cancer is a substantial source of morbidity and mortality worldwide. Estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor 2 (HER2), are the primary biomarkers which inform breast cancer treatment. Although endocrine therapy for ER+ patients is widely available, there is a need for increased access to low cost, rapid and accurate ER testing methods. In this work, we designed a near-infrared optical nanosensor using single-walled carbon nanotubes (SWCNT) as the transducer and an anti-ERα antibody as the recognition element. We evaluated the sensor in vitro prior to testing with 26 breast cancer samples which were collected by scraping the cut surface of fresh, surgically resected tumors. 20 samples were ER+, and 6 ER-, representing 13 unique patients. We found the nanosensor can differentiate ER- from ER+ patient biopsies through a shift in its center wavelength upon sample addition. Receiver operating characteristic area under the curve analyses determined that the strongest classifier with an AUC of 0.94 was the (7,5) SWCNT after direct incubation and measurement, and without further processing. We anticipate that further testing and development of this nanosensor may push its utility toward field-deployable, rapid ER subtyping with potential for additional molecular marker profiling.