Intratumoral heterogeneity of second-harmonic generation scattering from tumor collagen and its effects on metastatic risk prediction

Abstract

Background: Metastases are the leading cause of breast cancer-related deaths. The tumor microenvironment impacts cancer progression and metastatic ability. Fibrillar collagen, a major extracellular matrix component, can be studied using the light scattering phenomenon known as second-harmonic generation (SHG). The ratio of forward- to backward-scattered SHG photons (F/B) is sensitive to collagen fiber internal structure and has been shown to be an independent prognostic indicator of metastasis-free survival time (MFS). Here we assess the effects of heterogeneity in the tumor matrix on the possible use of F/B as a prognostic tool.

Methods: SHG imaging was performed on sectioned primary tumor excisions from 95 untreated, estrogen receptor-positive, lymph node negative invasive ductal carcinoma patients. We identified two distinct regions whose collagen displayed different average F/B values, indicative of spatial heterogeneity: the cellular tumor bulk and surrounding tumor-stroma interface. To evaluate the impact of heterogeneity on F/B's prognostic ability, we performed SHG imaging in the tumor bulk and tumor-stroma interface, calculated a 21-gene recurrence score (surrogate for OncotypeDX®, or S-ODX) for each patient and evaluated their combined prognostic ability.

Results: We found that F/B measured in tumor-stroma interface, but not tumor bulk, is prognostic of MFS using three methods to select pixels for analysis: an intensity threshold selected by a blinded observer, a histogram-based thresholding method, and an adaptive thresholding method. Using both regression trees and Random Survival Forests for MFS outcome, we obtained data-driven prediction rules that show F/B from tumor-stroma interface, but not tumor bulk, and S-ODX both contribute to predicting MFS in this patient cohort. We also separated patients into low-intermediate (S-ODX < 26) and high risk (S-ODX ≥26) groups. In the low-intermediate risk group, comprised of patients not typically recommended for adjuvant chemotherapy, we find that F/B from the tumor-stroma interface is prognostic of MFS and can identify a patient cohort with poor outcomes.

Conclusions: These data demonstrate that intratumoral heterogeneity in F/B values can play an important role in its possible use as a prognostic marker, and that F/B from tumor-stroma interface of primary tumor excisions may provide useful information to stratify patients by metastatic risk.

Keywords: Breast cancer; Collagen; F/B; Metastasis; Multiphoton microscopy; Prognosis; Second-harmonic generation; Tumor microenvironment.

Fig. 1

Collagen F/B is heterogeneous within primary breast tumor tissue. SHG F/B images (a series of adjacent ROIs extending along the x-axis) and matching H&E images were stitched end-to-end to form a composite ROI for a representative sample. Examples of tumor bulk (solid box), tumor-stroma interface (dashed box), and uninvolved tissue (circle) are shown. The F/B values were plotted in a false color “heatmap” (low F/B values = dark blue, high F/B values = yellow) to illustrate differences in F/B within the tumor region including tumor bulk and tumor-stroma interface, indicating intratumoral heterogeneity in F/B

Fig. 2

Collagen features vary between the tumor bulk and the tumor-stroma interface of primary tumor excisions. Primary tumor excisions contain both tumor bulk (solid boxes) and tumor-stroma interface (dashed boxes). Tumor bulk consists of tumor cell clusters surrounded by individual SHG-producing fiber bundles. The tumor-stroma interface is comprised mainly of closely packed collagen fibers and individual stromal cells adjacent to the tumor bulk. Representative SHG F/B and matching H&E images from 3 individual patients are shown. The scale bar applies to all images in this figure

Fig. 3

F/B in tumor-stroma interface versus tumor bulk. Collagen fiber internal structure, as represented by F/B, is significantly different between tumor bulk and the tumor-stroma interface in IDC ER+ LNN excised primary tumors. Error bars = SD, t-test, p < 0.0001, n = 92

Fig. 4

F/B measured in tumor-stroma interface and bulk of primary tumor sections and relation to MFS. SHG F/B values were produced by a user-defined threshold for each individual image from the a, tumor bulk and b, tumor-stroma interface. Patients were split into four equal quartiles (Q1 = lowest F/B) based on F/B, and the percentage of each quartile surviving without metastasis then plotted versus time. Tick marks represent censoring events caused when a patient dies of a cause other than cancer or is lost to follow-up. Partial likelihood ratio tests for ln F/B: p = 0.05 (tumor bulk, n = 95) and p = 0.00008 (tumor-stroma interface, n = 92)

Fig. 5

F/B generated using histogram-based thresholding and its relation to MFS. SHG F/B values were produced using binary masks generated by a histogram-based thresholding method for each individual image taken in the a, tumor bulk and b, tumor-stroma interface. Patients were split into four equal quartiles (Q1 = lowest F/B) based on F/B, and the percentage of each quartile surviving without metastasis then plotted versus time. Tick marks represent censoring events caused when a patient dies of a cause other than cancer or is lost to follow-up. Partial likelihood ratio tests for ln F/B: p = 0.01 (tumor bulk, n = 95) and p = 0.0009 (tumor-stroma interface, n = 92)

Fig. 6

F/B generated using adaptive thresholding and its relation to MFS. An adaptive thresholding method was used to create binary masks used to calculate SHG F/B in the a, tumor bulk and b, tumor-stroma interface. Patients were split into four equal quartiles (Q1 = lowest F/B) based on F/B, and the percentage of each quartile surviving without metastasis then plotted versus time. Tick marks represent censoring events caused when a patient dies of a cause other than cancer or is lost to follow-up. Partial likelihood ratio test p = 0.4 (tumor bulk, n = 95) and p = 0.002 (tumor-stroma interface, n = 92)

Fig. 7

Regression tree derived using the method of Leblanc & Crowley. When given the SODX score and all six methods of generating F/B as inputs, this algorithm selects F/B from the tumor-stroma interface, calculated using the adaptive thresholding method (“FB_THI_AT”), and SODX score (“SODX_score”), as predictors of MFS. The RSF method (results not shown) identifies the same two predictors has having the highest variable importance in predicting MFS

Fig. 8

F/B is prognostic of MFS in the tumor-stroma interface of primary excisions with S-ODX < 26. Patients were divided by S-ODX score and then each group was split into four equal quartiles (Q1 = lowest F/B) based on F/B. The three higher quartiles (Q2-Q4) were pooled, and the percentage of each quartile surviving without metastasis then plotted versus time for these two groups (Q1 and Q2-Q4). a, S-ODX < 26 and b, S-ODX ≥26. Tick marks represent censoring events caused when a patient dies of a cause other than cancer or is lost to follow-up. Partial likelihood ratio test a p = 0.008, n = 62, b p = 0.4, n = 30