Abstract B86: Novel assessment of SPARC expression by hierarchical clustering in pancreatic ductal adenocarcinoma shows distinct prognostic and predictive groups
Bibliographic record
Abstract
Abstract Background: Secreted Protein, Acid, Cysteine-Rich (SPARC) has been classified as a marker of poor prognosis and has been postulated as a therapeutic target in pancreatic ductal adenocarcinoma (PDAC). The clinical trial findings have been discordant with Phase I/II trials declaring that patients treated with nab-paclitaxel and gemcitabine whose tumors express SPARC in the stromal component have a significantly improved response. Unfortunately, these results were not validated in a subsequent Phase III trial. The focus of examining SPARC expression in a particular component and using nab-paclitaxel, which binds to SPARC, sets the stage for partial clinical efficacy due to component expression heterogeneity. The goal of this study is to develop an integrated component based approach to the quantification of SPARC in PDAC to reveal discrete prognostic and predictive groups in a cohort of resected patients treated with an adjuvant pyrimidine analog or subjected to post-surgical observation. Methods: Tissue-microarray based immunohistochemical quantification of SPARC expression was performed in the epithelial and stromal histological components of 246 resected PDACs from the Vancouver Coastal Health Region collected between 1989-2013. The semi-quantitative H-Score methodology was used where the percent of cells staining for SPARC is multiplied by the subjective assessment of its intensity (1-3) resulting in a range of scores between 0 - 300. These scores were subjected to unsupervised hierarchical clustering using Ward’s algorithm. The number of clusters was determined through an a priori decision that no cluster could have a resultant N < 30. Comparisons of component specific H-Scores across the resultant clusters were performed with the non-parametric Steel-Dwass test for multiple comparisons. Univariable disease specific survival (DSS) analysis was performed with the Kaplan-Meier method to examine the cluster specific survival profiles with regard to prognostic and predictive effects. The proportional hazards model was used to perform multivariable DSS to determine if the resultant prognostic groups were statistically independent when other known prognostic variables were considered. Results: The distribution of H-Scores spanned the entire range of possible values for both the epithelial and stromal components with medians [IQRs] of 120 [120] and 270 [120] respectively. The two-variable hierarchical clustering procedure yielded six clusters ranging in size from 31 to 51 cases. Kaplan-Meier curves showed no statistically significant differences between any of the clusters. However, the cluster exhibiting the best prognosis (N=36) also had significantly lower H-Scores for the stromal component compared to the remaining 5 clusters (p<0.0001). Similar differences were observed for the epithelial component with statistically significant differences observed in 4 (p<=0.0008) of the 5 comparisons. Based on these data, the five clusters with higher expression and one with low were respectively categorized as SPARCHigh (N =243) and SPARCLow (N=36). Kaplan-Meier analysis showed that the SPARCLow cohort had a 7.4-month survival advantage (p=0.03). Multivariable analysis was performed including: age at surgery, adjuvant chemotherapy, lymphovascular invasion, perineural invasion, pN Stage, and tumor budding which showed that the SPARCLow/High categorization is of independent prognostic significance (p=0.03). Assessment of the predictive ability of the six SPARC clusters showed that three of the six clusters representing 117 (47%) patients in this cohort would benefit from adjuvant pyrimidine analog based therapy. The three predictive clusters had concordant high or low expression in both components. Conclusion: This study illustrates that there is enhanced value in a combinatorial approach to the examination of SPARC in the stromal and epithelial components of PDAC where we have confirmed its status as a negative prognostic marker. While this cohort cannot address the issue of nab-paclitaxel in PDAC, we have noted that heterogeneity of SPARC expression at the component level can yield discrete predictive groups with regard to response to adjuvant pyrimidine analogs. Citation Format: Steve E. Kalloger, Joanna M. Karasinska, HuiLi Wong, Daniel J. Renouf, David F. Schaeffer.{Authors}. Novel assessment of SPARC expression by hierarchical clustering in pancreatic ductal adenocarcinoma shows distinct prognostic and predictive groups. [abstract]. In: Proceedings of the AACR Special Conference on Pancreatic Cancer: Advances in Science and Clinical Care; 2016 May 12-15; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2016;76(24 Suppl):Abstract nr B86.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".