Collagen Fibers Structure Detected By Second-Harmonic Generation Microscopy: A Potential Prognostic Marker for High-Risk Nodular Sclerosis

Introduction: Second-Harmonic Generation microscopy (SHG) has provided progresses in extracellular matrix research, mainly regarding automated analysis of collagen fibers. Nodular sclerosis-classical Hodgkin lymphoma (NS) often has a rich collagen deposition, which remains poorly explored in its biological significance and potential prognostic role. The aim of this study was to characterize the collagen component of NS using SHG, and to investigate its clinical value.

Methods: Hematoxylin and eosin stained slides from 53 consecutive samples of paraffin embedded NS tissue were analyzed by SHG imaging in an Inverted Zeiss LSM 780-NLO. HIV-positive individuals were excluded. For comparative purposes, 11 reactive lymph nodes (RL) were also randomly selected. In each slide, 3 capsular areas and 3 sclerotic regions (perinodular septa - PS - in NS and fibrotic foci around germinal centers in RL) were chosen. Collagen near blood vessels was not considered. We evaluated quantity, uniformity and organization of the fibers with ImageJ and OrientationJ plug-in in 4 hotspots from each image. Shapiro-Wilk test was used to assess data normality, while t-tests and Pearson correlations were performed to analyze collagen parameters between two groups. Overall survival (OS) was defined as time from diagnosis until death from the disease or last follow-up. Event-free survival (EFS) was set as time from diagnosis until progressive disease, death from the disease or last follow-up. Collagen data were used in survival analyses as continuous variables (in Cox-hazards model) or categorical ones (by choosing the mean value as a threshold for Kaplan-Meier curves and log-rank test). Significance was set at p<0.05.

Results: NS patients were mostly female (58.5%). The mean and median values of age at diagnosis were, respectively, 33 and 29 years (range: 14-82 years). Forty-two patients (79.2%) had B-symptoms, while bulky disease was seen in 23 cases (43.3%). Bone marrow was infiltrated in 4 patients (7.5%). Based on the International Prognostic Score (IPS), patients were stratified as follows: 39 (73.5%) as low-risk (IPS ≤ 3) and 14 (26.5%) as high-risk (IPS > 3). The first-line treatment was ABVD in 40 cases (75.5%) and BEACOPP in 13 patients (24.5%). Radiotherapy was performed in 27 (50.9%) patients.

Tumor PS presented more dense (p<0.01), organized (p<0.01) and uniform (p=0.02) fibers than sclerosis in RL. NS capsule had only a more organized collagen than RL (p=0.02). Considering solely NS, we found significant correlations of collagen fibers quantity (r=0.67), uniformity (r=0.78) and organization (r=0.49) between capsule and PS, suggesting a proportional deposition of collagen in these compartments. When clinical data were analyzed in the entire patient group, the presence of more organized PS collagen was associated with low-risk IPS (p=0.03), grade 2 tumors (p= 0.04), extranodal disease (p=0.003) and a higher risk of death (p=0.02). On the other hand, a higher fiber organization in NS capsule was only associated with low-risk IPS (p=0.03).

The presence of high-risk IPS was associated with a shorter EFS (p=0.002), but had no relationship to OS. Collagen parameters had no impact on EFS. However, in univariate Cox regression, higher collagen quantity (p=0.04) in PS was the only factor associated with a worse OS. When only high-risk IPS cases were analyzed in Cox regression model, PS collagen had association with unfavorable OS regarding higher collagen quantity (p=0.03) and uniformity (p<0.01). In this setting, a higher quantity of capsular collagen was also associated with a worse OS (p=0.02). We did not perform multivariate analysis in this group due to the small amount of patients (14). Also in the high-risk cases, using the mean value of PS collagen quantity (15.19) as a threshold to binarize data, we found a trend to worse OS in cases with a higher collagen quantity (p=0.05, figure 1). Considering only low-risk IPS cases, collagen factors had no association with OS.

Conclusion: Collagen parameters in NS have distinct features than in RL and affect clinical presentation of the tumor. Moreover, collagen quantity and uniformity in cases with high-risk IPS were associated with survival, which indicates that incorporation of collagen information might be relevant in the prognostication of NS.

Figure 1.

View largeDownload slide

Survival curve of high-risk patients (n=14) stratified according to the perinodular collagen (pCOL) quantity threshold.

Figure 1.

View largeDownload slide

Survival curve of high-risk patients (n=14) stratified according to the perinodular collagen (pCOL) quantity threshold.

Close modal