Early Assessment Window for Predicting Breast Cancer Neoadjuvant Therapy using Biomarkers, Ultrasound, and Diffuse Optical Tomography

Abstract

Purpose: The purpose of the study was to assess the utility of tumor biomarkers, ultrasound (US) and US-guided diffuse optical tomography (DOT) in early prediction of breast cancer response to neoadjuvant therapy (NAT).

Methods: This prospective HIPAA compliant study was approved by the institutional review board. Forty one patients were imaged with US and US-guided DOT prior to NAT, at completion of the first three treatment cycles, and prior to definitive surgery from February 2017 to January 2020. Miller-Payne grading was used to assess pathologic response. Receiver operating characteristic curves (ROCs) were derived from logistic regression using independent variables, including: tumor biomarkers, US maximum diameter, percentage reduction of the diameter (%US), pretreatment maximum total hemoglobin concentration (HbT) and percentage reduction in HbT (%HbT) at different treatment time points. Resulting ROCs were compared using area under the curve (AUC). Statistical significance was tested using two-sided two-sample student t-test with P < 0.05 considered statistically significant. Logistic regression was used for ROC analysis.

Results: Thirty-eight patients (mean age = 47, range 24-71 years) successfully completed the study, including 15 HER2 + of which 11 were ER + ; 12 ER + or PR + /HER2-, and 11 triple negative. The combination of HER2 and ER biomarkers, %HbT at the end of cycle 1 (EOC1) and %US (EOC1) provided the best early prediction, AUC = 0.941 (95% CI 0.869-1.0). Similarly an AUC of 0.910 (95% CI 0.810-1.0) with %US (EOC1) and %HbT (EOC1) can be achieved independent of HER2 and ER status. The most accurate prediction, AUC = 0.974 (95% CI 0.933-1.0), was achieved with %US at EOC1 and %HbT (EOC3) independent of biomarker status.

Conclusion: The combined use of tumor HER2 and ER status, US, and US-guided DOT may provide accurate prediction of NAT response as early as the completion of the first treatment cycle.

Clinical trial registration number: NCT02891681. https://clinicaltrials.gov/ct2/show/NCT02891681 , Registration time: September 7, 2016.

Keywords: Near Infrared imaging; Personalized medicine; Predicting neoadjuvant therapy; Ultrasound.

Fig. 1

Patient Study Flow Diagram. Exclusion criteria included pregnancy, breastfeeding, prior history of breast cancer, prior history of chest wall radiation, prior history of breast reconstruction, reduction, or augmentation and bilateral breast cancers

Fig. 2

US-guided DOT probe. The foot-print of the combined DOT probe is approximately 10 cm. Four source laser diodes of 730 nm, 785 nm, 808 nm and 830 nm optical wavelengths were sequentially switched to nine source positions (pointed by an arrow) on the probe, while the reflected light was coupled by the 14 light guides (pointed by an arrow) to 14 parallel detectors

Fig. 3

a HbT of Miller-Payne grade 4–5 tumors (therapy responders) and grade 1–3 tumors after 1, 2, 3 cycles of neoadjuvant therapy. The unit is μM. b %HbT of grade 4–5 tumors vs. grade 1–3 tumors after first three cycles of neoadjuvant therapy. c %US of grade 4–5 tumors vs. grade 1–3 tumors after first three cycles of neoadjuvant therapy. Pc is Bonferroni-Holm corrected P value

Fig. 4

A 59 year-old patient with a T3 triple negative cancer and treated with ACT. The US/DOT imaging were performed at baseline, end of cycle 1 (EOC1), 2, 3, 5 and before surgery. a–c are co-registered US images obtained at baseline, EOC1 and EOC3. The largest lesion diameters measured by US were 4.6 cm, 3.4 cm, 1.0 cm. The corresponding %US at EOC1 and EOC 3 were 73.9% and 21.7%. e–f are corresponding HbT maps. Each map has 7 slices reconstructed at depths from 0.5 cm to 3.5 cm with 0.5 cm spacing. Each slice has spatial dimensions of 9 cm by 9 cm. The maximum HbT measured at baseline, EOC1, and EOC3 were 108.7 μM, 73.5 μM, and 45.0 μM. The %HbT were 67.6% and 41.4% at EOC1 and EOC3. The patient achieved pCR with Miller-Payne of 5

Fig. 5

A 62 year-old patient with a T2 ER negative PR positive and HER2 negative IDC and treated with ACT. The US/DOT imaging were performed at baseline, end of cycle 1, 2, 3 and before surgery. a–c are co-registered US images obtained at baseline, EOC1 and EOC3. The largest lesion diameters measured by US were 3.6 cm, 2.4 cm, 1.7 cm. The corresponding %US at EOC1 and EOC 3 were 66.7% and 47.2%. e–f are corresponding HbT maps. Each map has 7 slices reconstructed at depths from 0.5 cm to 3.5 cm with 0.5 cm spacing. Each slice has a spatial dimensions of 9 cm by 9 cm. The maximum HbT measured at baseline, EOC1, and EOC3 were 70.3 μM, 79.3 μM, and 65.8 μm. The %HbT were 112.8% and 93.6% at EOC1 and EOC3. The patient had 2.4 cm residual tumor with no histologic evidence of tumor response as evaluated after the surgery. Miller-Payne grade was 1

Fig. 6

ROCs obtained from different set of predictor variables. a ROCs of known HER2/ER subgroup with 5 sets of predictor variables, and b ROCs based on HbT, %US-EOC1, %HbT and %US changes regardless of biomarkers