Molecular Stratification Research Articles

Abstract 5302: Molecular subtype stratification for prostate cancer from mpMRI and histopathology images using convolutional neural networks and transfer learning

Abstract Objective: Molecular subtypes have been found to be associated with prognosis and treatment response for prostate cancer (PCa) patients and have the potential to aid personalized treatment planning. We aim to stratify molecular subtypes by lesion characterization from multiparametric magnetic resonance images (mp-MRI) using convolutional neural networks (CNN) and knowledge transferred from lesion malignancy classification task. Methods: We identified 23 PCa patients with available mpMRI and molecular subtype information. Each patient may harbor multi-focal lesions of different molecular subtypes. Automated antibody based dual-color immunohistochemistry assays were developed for the simultaneous assessment of ERG-PTEN and ERG-SPINK1 status in PCa on the whole mount microscopic sections. In 7 patients 11 intraprostatic lesions (ILs) were identified as ERG+, 3 patients 3 ILs as SPINK1+, and 10 patients 11 ILs as triple negative (ERG-, SPINK1- and ETS-). We fed a CNN using intratumoral region of interest defined in a slice-by-slice manner from T2WI, ADC and DWIb50. The feature maps from the third convolutional layer were flattened into 7,744 dimensional vectors to represent each IL slice. To compensate for the small sample size, we utilized transfer learning, from the task of tumor malignancy stratification. The CNN model was pre-trained on mpMRI with 320 ILs from 201 patients for malignancy stratification using a different cohort we published previously, which was later fined tuned on this cohort on malignancy classification for domain adaptation purpose. Results: The clustering accuracy (Top 3 estimates) using cosine similarity metrics was shown in Table 1 for each molecular subtype category respectively. The preliminary results supported our hypothesis that task of lesion malignancy and molecular sub-types stratification were correlated in the imaging features derived from mpMRI. Table 1:Clustering accuracy for each molecular category. Each ILs input had 3 consecutive slices in a sequence, know as 2.5D, to incorporate lesion growth pattern.Accuracy (# of 2.5D image slices)Top 1 predictionsTop 2 predictionsTop 3 predictionsERG+ (21)0.670.900.95SPINK1+ (4)0.50.750.75Negative (17)0.650.760.76 Conclusions: This work showed the potential to classify the molecular subtypes of PCa from mp-MRI. The small sample size problem was tackled using transfer learning. Acknowledgement: The work was supported by a Research Scholar Grant: RSG-15-137-01-CCE from the American Cancer Society. Citation Format: Weiwei Zong, Eric N. Carver, Aharon Feldman, Nallasivam Palanisamy, Ning Wen. Molecular subtype stratification for prostate cancer from mpMRI and histopathology images using convolutional neural networks and transfer learning [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5302.

A gene expression-based single sample predictor of lung adenocarcinoma molecular subtype and prognosis.

Disease recurrence in surgically treated lung adenocarcinoma (AC) remains high. New approaches for risk stratification beyond tumor stage are needed. Gene expression‐based AC subtypes such as the Cancer Genome Atlas Network (TCGA) terminal‐respiratory unit (TRU), proximal‐inflammatory (PI) and proximal‐proliferative (PP) subtypes have been associated with prognosis, but show methodological limitations for robust clinical use. We aimed to derive a platform independent single sample predictor (SSP) for molecular subtype assignment and risk stratification that could function in a clinical setting. Two‐class (TRU/nonTRU=SSP2) and three‐class (TRU/PP/PI=SSP3) SSPs using the AIMS algorithm were trained in 1655 ACs (n = 9659 genes) from public repositories vs TCGA centroid subtypes. Validation and survival analysis were performed in 977 patients using overall survival (OS) and distant metastasis‐free survival (DMFS) as endpoints. In the validation cohort, SSP2 and SSP3 showed accuracies of 0.85 and 0.81, respectively. SSPs captured relevant biology previously associated with the TCGA subtypes and were associated with prognosis. In survival analysis, OS and DMFS for cases discordantly classified between TCGA and SSP2 favored the SSP2 classification. In resected Stage I patients, SSP2 identified TRU‐cases with better OS (hazard ratio [HR] = 0.30; 95% confidence interval [CI] = 0.18‐0.49) and DMFS (TRU HR = 0.52; 95% CI = 0.33‐0.83) independent of age, Stage IA/IB and gender. SSP2 was transformed into a NanoString nCounter assay and tested in 44 Stage I patients using RNA from formalin‐fixed tissue, providing prognostic stratification (relapse‐free interval, HR = 3.2; 95% CI = 1.2‐8.8). In conclusion, gene expression‐based SSPs can provide molecular subtype and independent prognostic information in early‐stage lung ACs. SSPs may overcome critical limitations in the applicability of gene signatures in lung cancer.

Gene Expression Clustering and Selected Head and Neck Cancer Gene Signatures Highlight Risk Probability Differences in Oral Premalignant Lesions.

Background: Oral premalignant lesions (OPLs) represent the most common oral precancerous conditions. One of the major challenges in this field is the identification of OPLs at higher risk for oral squamous cell cancer (OSCC) development, by discovering molecular pathways deregulated in the early steps of malignant transformation. Analysis of deregulated levels of single genes and pathways has been successfully applied to head and neck squamous cell cancers (HNSCC) and OSCC with prognostic/predictive implications. Exploiting the availability of gene expression profile and clinical follow-up information of a well-characterized cohort of OPL patients, we aim to dissect tissue OPL gene expression to identify molecular clusters/signatures associated with oral cancer free survival (OCFS). Materials and methods: The gene expression data of 86 OPL patients were challenged with: an HNSCC specific 6 molecular subtypes model (Immune related: HPV related, Defense Response and Immunoreactive; Mesenchymal, Hypoxia and Classical); one OSCC-specific signature (13 genes); two metabolism-related signatures (3 genes and signatures raised from 6 metabolic pathways associated with prognosis in HNSCC and OSCC, respectively); a hypoxia gene signature. The molecular stratification and high versus low expression of the signatures were correlated with OCFS by Kaplan–Meier analyses. The association of gene expression profiles among the tested biological models and clinical covariates was tested through variance partition analysis. Results: Patients with Mesenchymal, Hypoxia and Classical clusters showed an higher risk of malignant transformation in comparison with immune-related ones (log-rank test, p = 0.0052) and they expressed four enriched hallmarks: “TGF beta signaling” “angiogenesis”, “unfolded protein response”, “apical junction”. Overall, 54 cases entered in the immune related clusters, while the remaining 32 cases belonged to the other clusters. No other signatures showed association with OCFS. Our variance partition analysis proved that clinical and molecular features are able to explain only 21% of gene expression data variability, while the remaining 79% refers to residuals independent of known parameters. Conclusions: Applying the existing signatures derived from HNSCC to OPL, we identified only a protective effect for immune-related signatures. Other gene expression profiles derived from overt cancers were not able to identify the risk of malignant transformation, possibly because they are linked to later stages of cancer progression. The availability of a new well-characterized set of OPL patients and further research is needed to improve the identification of adequate prognosticators in OPLs.

Whole brain radiotherapy (WBRT) for leptomeningeal metastasis from NSCLC in the era of targeted therapy: a retrospective study

Background and purposeLeptomeningeal metastasis (LM) is a rare but detrimental complication in patients with non-small cell lung cancer (NSCLC). Although whole brain radiotherapy (WBRT) is used to eliminating cancer cells or microscopic foci, it is becoming less favorable due to the concerns over neurocognitive toxicity. This study aimed to re-evaluate the role of WBRT in the setting of modern targeted therapy.Materials and methodsFrom December 2014 to March 2019, 80 NSCLC patients with cytologically and/or radiologically proven LM diagnosis were retrospectively analyzed.ResultsThe median OS (mOS) after diagnosis of LM was 8.0 (95%CI: 4.4 to 11.6) months, and the one-year OS was 39.4%. The mOS for EGFR-mutated LM patients was 12.6 (3.0 to 22.2) months versus only 4.1 (2.8 to 5.4) for patients with wild-type EGFR (P < 0.001). Younger patients (< 53.5 yrs.) appeared to have a better OS than older patients (≥53.5 yrs.) (12.6 vs. 6.1, P = 0.041). No survival benefits were found in EGFR-mutated patients who received WBRT (P = 0.490). In contrast, mOS was significantly prolonged in wild-type EGFR patients with WBRT versus non-WBRT (mOS: 8.0 vs. 2.1, P = 0.002). Multivariate analysis indicated that WBRT (P = 0.025) and younger age (P = 0.048) were independent prognostic factors that predicted prolonged survival for wild-type EGFR LM patients from NSCLC.ConclusionOur study demonstrated that WBRT has clear survival advantages for patients with wild-type EGFR, and molecular biological stratification of LM patients for WBRT is highly recommended.

Phase II Study of Arginine Deprivation Therapy With Pegargiminase in Patients With Relapsed Sensitive or Refractory Small-cell Lung Cancer

BackgroundPre-clinical studies indicated that arginine-deprivation therapy using pegylated arginine deiminase (pegargiminase, ADI-PEG 20) may be effective in patients with argininosuccinate synthetase 1 (ASS1)-deficient small-cell lung cancer (SCLC). Patients and MethodsPatients were enrolled into either a ‘sensitive’ disease cohort (≥ 90 days response to first-line chemotherapy) or a ‘refractory’ disease cohort (progression while on chemotherapy or < 90 days afterwards or ≥ third-line treatment). Patients received weekly intramuscular pegargiminase, 320 IU/m2 (36.8 mg/m2), until unacceptable toxicity or disease progression. The primary endpoint was tumor response assessed by Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 with secondary endpoints including tolerability, pharmacodynamics, and immunogenicity. ResultsBetween January 2011 and January 2014, 22 patients were enrolled: 9 in the sensitive disease cohort and 13 in the refractory disease cohort. At a pre-planned interim analysis, the best overall response observed was stable disease in 2 patients in each cohort (18.2%). Owing to the lack of response and slow accrual in the sensitive disease cohort, the study was terminated early. Pegargiminase treatment was well-tolerated with no unexpected adverse events or discontinuations. ConclusionAlthough pegargiminase monotherapy in SCLC failed to meet its primary endpoint of RECIST-confirmed responses, more recent molecular stratification, including MYC status, may provide new opportunities moving forward.

A Summary of the 2020 Gastric Cancer Summit at Stanford University

There exists no coherent national strategy for the early detection or prevention of gastric cancer in the United States (US), even among identified high-risk groups such as Asian Americans, African Americans, Hispanic Americans, and Alaska Native/American Indian peoples. As a result, patients with gastric cancer in the US are diagnosed at later stages and demonstrate worse overall survival compared to nations of East Asia with established screening programs (Table 1). The under-recognition of gastric cancer risk within minority communities is a significant unaddressed healthcare disparity.

From the Editor's Desk…

From the Editor's Desk…

A Molecular Stratification of Chilean Gastric Cancer Patients with Potential Clinical Applicability.

Gastric cancer (GC) is a complex and heterogeneous disease. In recent decades, The Cancer Genome Atlas (TCGA) and the Asian Cancer Research Group (ACRG) defined GC molecular subtypes. Unfortunately, these systems require high-cost and complex techniques and consequently their impact in the clinic has remained limited. Additionally, most of these studies are based on European, Asian, or North American GC cohorts. Herein, we report a molecular classification of Chilean GC patients into five subtypes, based on immunohistochemical (IHC) and in situ hybridization (ISH) methods. These were Epstein–Barr virus positive (EBV+), mismatch repair-deficient (MMR-D), epithelial to mesenchymal transition (EMT)-like, and accumulated (p53+) or undetected p53 (p53−). Given its lower costs this system has the potential for clinical applicability. Our results confirm relevant molecular alterations previously reported by TCGA and ACRG. We confirm EBV+ and MMR-D patients had the best prognosis and could be candidates for immunotherapy. Conversely, EMT-like displayed the poorest prognosis; our data suggest FGFR2 or KRAS could serve as potential actionable targets for these patients. Finally, we propose a low-cost step-by-step stratification system for GC patients. To the best of our knowledge, this is the first Latin American report on a molecular classification for GC. Pending further validation, this stratification system could be implemented into the routine clinic

Oncogenic Mutations in Tumorigenesis and Targeted Therapy in Breast Cancer

Breast cancer is the most frequently identified cancer and the leading cause of women death from cancer worldwide. Although, based on the molecular classification, various therapeutic interventions are available for this deadly disease. However, increasing evidence suggests existence of differential mutation patterns in various cellular pathways crucial for the regulation of tumor microenvironment, cellular plasticity, and genomic instability. Genomic instability leads to the deregulation of oncogenic signaling pathways and makes it harder to find specific target in order to apply an effective therapy. Additionally, the deregulated oncogenic pathways acquire the gain of functions which makes the tumor more resilient to the available therapeutic intervention. These oncogenic mutations are also responsible for aiding up to tumor heterogeneity. Tumor heterogeneity in turn increases the intra-tumor diversity and the existence of various clonal population complicates the tumor diagnostic which leads to therapeutic challenges. In this review, we talked about the need of further molecular stratification of breast cancer to better understand the targetable mutation pattern and to design the effective outcome of treatment. Tumor heterogeneity has a serious clinical consequence for the breast cancer patients. However, there are no known ways to target the tumor heterogeneity yet. Although, we think that the future is in the hand of personalized medication in order to target the tumor heterogeneity, which still has a long way to go. Current treatment choices are mostly based on the molecular classification such as status of hormone receptors estrogen receptor, progesterone receptor, and HER2 expressions. The advancement in the global sequencing devices has given great advantages in understanding the tumor heterogeneity and identification of additional mutation patterns, which are being targeted in combination with available treatments for the better outcome of the patient’s survival. We summarize the clinical challenges and opportunities for better management of breast cancer patients and treatment choices. Investigating the mutational signatures and their associated mechanisms responsible for cancer could be a new evolving target and thus improve the opportunities for therapy.

ProstaTrend—A Multivariable Prognostic RNA Expression Score for Aggressive Prostate Cancer

BackgroundProstate cancer (PCa) is the most prevalent solid cancer among men in Western Countries. The clinical behavior of localized PCa is highly variable. Some cancers are aggressive leading to death, while others can even be monitored safely. Hence, there is a high clinical need for precise biomarkers for identification of aggressive disease in addition to established clinical parameters. ObjectiveTo develop an RNA expression-based score for the prediction of PCa prognosis that facilitates clinical decision making. Design, setting, and participantsWe assessed 233 tissue specimens of PCa patients with long-term follow-up data from fresh-frozen radical prostatectomies (RPs), from formalin-fixed and paraffin-embedded RP specimens and biopsies by transcriptome-wide next-generation sequencing and customized expression microarrays. Outcome measurements and statistical analysisWe applied Cox proportional hazard models to the cohorts from different platforms and specimen types. Evidence from these models was combined by fixed-effect meta-analysis to identify genes predictive of the time to death of disease (DoD). Genes were combined by a weighted median approach into a prognostic score called ProstaTrend and transferred for the prediction of biochemical recurrence (BCR) after RP in an independent cohort of The Cancer Genome Atlas (TCGA). Results and limitationsProstaTrend comprising ∼1400 genes was significantly associated with DoD in the training cohort of PCa patients treated by RP (leave-one-out cross-validation, Cox regression: p=2e-09) and with BCR in the TCGA validation cohort (Cox regression: p=3e-06). The prognostic impact persisted after multivariable Cox regression analysis adjusting for Gleason grading group (GG) ≥3 and resection status (p=0.001; DoD, training cohort) and for GG≥3, pathological stage ≥T3, and resection state (p=0.037; BCR, validation cohort). ConclusionsProstaTrend is a transcriptome-based score that predicts DoD and BCR in cohorts of PCa patients treated with RP. Patient summaryProstaTrend provides molecular patient risk stratification after radical prostatectomy.

Abstract B20: Endometrial cancer molecular risk stratification in endometrioid ovarian cancers: A novel application of precision medicine

Abstract Objective: Endometrioid ovarian carcinoma (ENOC) is associated with a generally more favorable prognosis compared to other ovarian carcinoma histotypes. Nonetheless, patients are still treated with a “one size fits all” approach. While tumor staging offers some stratification, the development of personalized treatment concepts remains elusive. Our group has recently validated the Proactive Molecular Risk Classifier for Endometrial Cancer (ProMisE), to distinguish clinically relevant prognostic groups amongst endometrial carcinomas. ENOC shares risk factors, genomics, and histology with its endometrial counterpart. The aim of our study was to apply and test ProMisE on ENOC. Methods: ProMisE was applied to 509 ENOC after biomarker-assisted review of ENOC histotype. Cases were aligned into low-risk POLE mutant (POLE), moderate-risk mismatch repair deficient (MMRd), high-risk p53 abnormal (p53abn), and a final moderate-risk category lacking these biomarkers (p53wt). Kaplan-Meier and multivariate survival analysis were performed. Results: 4% of cases were POLE, 16% MMRd, 71% p53wt, and 10% p53abn. Groups showed distinct progression-free and overall survival (p &amp;lt;0.001), near identical to profiles of endometrial cancers. 5-year PFS was 54% in p53abn, 82% in MMRd, 85% in p53wt, and 100% in POLEmut cases. Median overall survival was reached in p53 mutant cases only. ProMisE classes of ENOC were independent of stage and residual disease in multivariable analysis. Conclusion: ProMisE risk classification provides additional prognostic information in a large cohort of ENOC. Our findings support introduction of ProMisE-stratified treatment strategies to improve patient care across ENOC. Further, ENOC may benefit from parallel efforts under investigation in endometrial carcinoma. Citation Format: Pauline Krämer, Aline Talhouk, Tjalling Bosse, Florian Heitz, Naveena Singh, Felix Kommoss, Sara Brucker, Jessica McAlpine, Stefan Kommoss, Martin Koebel, Michael Anglesio. Endometrial cancer molecular risk stratification in endometrioid ovarian cancers: A novel application of precision medicine [abstract]. In: Proceedings of the AACR Special Conference on Advances in Ovarian Cancer Research; 2019 Sep 13-16, 2019; Atlanta, GA. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(13_Suppl):Abstract nr B20.

Abstract A03: PrOTYPE (Predictor of high-grade-serous Ovarian carcinoma molecular subTYPE): The development and validation of a clinical-grade consensus classifier for the molecular subtypes of high-grade serous tubo-ovarian cancer

Abstract Background: Gene expression-based molecular subtypes of high-grade serous tubo-ovarian cancer (HGSOC) are distinguished by differential immune and stromal infiltration and may provide opportunities for targeted therapies. Integration of molecular subtypes into clinical trials has been hindered by inconsistent subtyping methodology. Methods: Adopting two independent approaches, we derived and internally validated algorithms for molecular subtype prediction from gene-expression array data in 1,650 tumors. We applied resulting models to assign labels to 3829 HGSOCs from the Ovarian Tumor Tissue Analysis (OTTA) consortium evaluated on NanoString. Using the labeled NanoString data, we developed, confirmed, and validated a minimal gene set, clinical-grade test and prediction tool. We also used the OTTA dataset to evaluate associations between molecular subtype, biologic, and clinical features. Findings: The locked-down test included a model with 55 genes that predicted HGSOC molecular subtype with &amp;gt;95% accuracy. Subtype varied between primary and metastatic site taken at the time of primary surgery, and was significantly associated with age, stage, CD8+ lymphocyte infiltration, residual disease, and outcome. In multivariable models, molecular subtypes lose their prognostic significance in the presence of risk factors such as residual disease and BRCA1/2 germline mutations. Interpretation: We validated the Predictor of high-grade-serous Ovarian carcinoma molecular subTYPE, or PrOTYPE, following the Institute of Medicine guidelines for the development of omics-based tests. This simple-to-use, cost-effective, fully defined, and locked-down clinical-grade assay will facilitate molecular subtype stratification into clinical trial design. PrOTYPE will allow for objective assessment of HGSOC molecular subtype predictive value in precision medicine applications. Citation Format: Aline Talhouk, Joshy George, Chen Wang, Ellen Goode, Susan Ramus, Jennifer Doherty, David Bowtell, Michael Anglesio, OTTA Consortium. PrOTYPE (Predictor of high-grade-serous Ovarian carcinoma molecular subTYPE): The development and validation of a clinical-grade consensus classifier for the molecular subtypes of high-grade serous tubo-ovarian cancer [abstract]. In: Proceedings of the AACR Special Conference on Advances in Ovarian Cancer Research; 2019 Sep 13-16, 2019; Atlanta, GA. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(13_Suppl):Abstract nr A03.

Differential metabolic profile associated with the condition of normoalbuminuria in the hypertensive population

Background and aimAlbuminuria is an indicator of sub-clinical organ damage and a marker of cardiovascular risk and renal disease. A percentage of hypertensive patients develop albuminuria despite being under chronic suppression of the renin-angiotensin system (RAS). We previously identified urinary metabolites associated with the development of albuminuria. In this study, we searched for metabolic alterations which reflect different levels within the condition of normoalbuminuria. Patients, materials and methodsUrine from 48 hypertensive patients under chronic RAS suppression was analyzed. They were classified according to the albumin / creatinine ratio (ACR) into 3 groups: Normoalbuminuria (<10 mg / g); Normal-high (10–30 mg / g in men, or 20−40 mg / g in women); and moderately high albuminuria (microalbuminuria, 30−200 mg / g or 40−300 mg / g, respectively). The metabolome was analyzed by mass spectrometry and a correlation analysis was performed between altered metabolite levels and ACR. ResultsOxaloacetate, 3-ureidopropionate, guanidoacetate and malate show significant variation between the normo and micro groups. Additionally, these metabolites are able to differentiate between patients in the normo and high-normal range. A significant correlation between metabolites and ACR was found. Observed variations point to alterations in the energy metabolism already in patients with albuminuria in the high-normal range. ConclusionsThe association between the molecular panel consisting of 3-ureidopropionate, oxaloacetate, malate and guanidoacetate and different levels of albuminuria is confirmed. A metabolic fingerprint was also identified showing variations within the condition of normoalbuminuria allowing an earlier molecular stratification of patients.

MRNA and microRNA selection for breast cancer molecular subtype stratification using meta-heuristic based algorithms

Cancer subtype stratification, which may help to make a better decision in treating cancerous patients, is one of the most crucial and challenging problems in cancer studies. To this end, various computational methods such as Feature selection, which enhances the accuracy of the classification and is an NP-Hard problem, have been proposed. However, the performance of the applied methods is still low and can be increased by the state-of-the-art and efficient methods. We used 11 efficient and popular meta-heuristic algorithms including WCC, LCA, GA, PSO, ACO, ICA, LA, HTS, FOA, DSOS and CUK along with SVM classifier to stratify human breast cancer molecular subtypes using mRNA and micro-RNA expression data. The applied algorithms select 186 mRNAs and 116 miRNAs out of 9692 mRNAs and 489 miRNAs, respectively. Although some of the selected mRNAs and miRNAs are common in different algorithms results, six miRNAs including miR-190b, miR-18a, miR-301a, miR-34c-5p, miR-18b, and miR-129-5p were selected by equal or more than three different algorithms. Further, six mRNAs, including HAUS6, LAMA2, TSPAN33, PLEKHM3, GFRA3, and DCBLD2, were chosen through two different algorithms. We have reported these miRNAs and mRNAs as important diagnostic biomarkers to the stratification of breast cancer subtypes. By investigating the literature, it is also observed that most of our reported mRNAs and miRNAs have been proposed and introduced as biomarkers in cancer subtypes stratification.

Adjuvant treatment of endometrial cancer in molecular era: Are we ready to move on?

For many decades, the Bokhman dualist vision was used to stratify endometrial cancer (EC) in good or bad tumors. Nowadays, a more robust and reliable molecular stratification is taking place with the The Cancer Genome Atlas Research Network (TCGA) classification bringing new and important information in the field. Collaborative groups are replicating TCGA using accessible tools with immunohistochemistry. It's time to move on and include this information along with pathology features to better delineate adjuvant treatment in EC.

Integrated transcriptomic and genomic analysis improves prediction of complete remission and survival in elderly patients with acute myeloid leukemia

Relevant molecular tools for treatment stratification of patients ≥65 years with acute myeloid leukemia (AML) are lacking. We combined clinical data with targeted DNA- and full RNA-sequencing of 182 intensively and palliatively treated patients to predict complete remission (CR) and survival in AML patients ≥65 years. Intensively treated patients with NPM1 and IDH2R172 mutations had longer overall survival (OS), whereas mutated TP53 conferred lower CR rates and shorter OS. FLT3-ITD and TP53 mutations predicted worse OS in palliatively treated patients. Gene expression levels most predictive of CR were combined with somatic mutations for an integrated risk stratification that we externally validated using the beatAML cohort. We defined a high-risk group with a CR rate of 20% in patients with mutated TP53, compared to 97% CR in low-risk patients defined by high expression of ZBTB7A and EEPD1 without TP53 mutations. Patients without these criteria had a CR rate of 54% (intermediate risk). The difference in CR rates translated into significant OS differences that outperformed ELN stratification for OS prediction. The results suggest that an integrated molecular risk stratification can improve prediction of CR and OS and could be used to guide treatment in elderly AML patients.

The potential of a CLIA-certified prognostic/predictive molecular test to address the rising costs of non-small cell lung cancer.

e21671 Background: Treating recurrences of non-small cell lung cancer (NSCLC) is increasingly expensive but still rarely curative. Disease free survival (DFS) among resected stage I-IIA patients remains only 50-70%. Guidelines advocate adjuvant therapy in “high-risk” patients in this population to reduce these costly and deadly recurrences, but recognize that conventional criteria have not been validated to stratify risk or predict benefit. A CLIA-certified, commercially available 14-gene expression risk profile (DetermaRx) has been extensively validated among stage I-IIA non-squamous NSCLC patients; prospective data now suggest that the test predicts improved DFS with adjuvant therapy. We therefore studied the potential economic impact of this molecular test on early stage NSCLC. Methods: Model variables included: relative increase in DFS with adjuvant treatment of molecular intermediate- and high-risk patients (25.5%; range 12.3-41.3%); cost of adjuvant ($8,760; $8,144-$9,376) or late-stage ($284,500; $224,900-345,200) treatment; and compliance with recommendations for adjuvant therapy (75%; 60-90%). Ranges were based on: historical trials and recent prospective data, published cost literature, and published survey data, respectively. Parameters were varied in a multifactorial, one-way sensitivity analysis to assess the impact of parameter uncertainty. Results: Reduction of recurrences with implementation of the 14-gene assay resulted in an average cost savings of $11,608/patient (potential systems savings of ~$450 million), even when including the cost of molecular risk stratification ($4000/patient) and of cisplatin-based adjuvant chemotherapy for molecular high- and intermediate-risk patients. Lower bound assumptions for relative improvement, cost of care, and compliance yielded persistent savings of $3,699, $8,091 and $8,486, respectively. Conclusions: Utilization of this predictive molecular risk stratification assay in the management of stage I-IIA non-squamous NSCLC has the potential to significantly reduce lung cancer costs in an era of targeted therapy and immunotherapy, while at the same time improving DFS and saving lives.

Exploring fingerprint ctDNA as a stratifying factor in colorectal cancer.

e16101 Background: Colorectal cancer (CRC) ranks near the top in tumor-related deaths. Its standard-of-care is surgical resection plus adjuvant chemotherapy, which can extend patient’s live and disease-free survival for up to several years. However, choosing an appropriate chemotherapy agent is difficult because CRC is a heterogenous disease. It is believed that molecular stratification can help to improve treatment accuracy and this has been proven by the consensus molecular subtypes (CMS) classification. However, the CMS classification relies on expensive and complex gene-expression profiling, and more cumbersomely, is limited to early stage patients without prior chemotherapy, radiation and metastasis. This situation denies many CRC patients from the benefit of CMS classification. Circulating tumor DNA (ctDNA) is a minimally invasive way to monitor disease progression and treatment response in solid tumors but its clinical utility in CRC remains to be validated. We explore the potential of using patient’s fingerprint ctDNA (defined below) as a stratifying factor in CRC patients to assist clinical decision. Methods: WES is performed on tumor and matched blood samples from 149 CRC patients. Based on the WES result, 20-30 tumor specific genes are selected for each patient and form their ctDNA fingerprints. The patients are grouped according to their level of fingerprint ctDNA (high- vs. low-ctDNA). Results: The two groups of patients show significant difference in treatment responses and mutational profiles. The low-ctDNA group in general respond to treatment well and have good prognosis. The high-ctDNA group, in contrast, often experience relapse or recurrence. Interestingly, the low-ctDNA group is dominated by point and small indel mutations in the top mutated genes while the high-ctDNA group is dominated by gene copy variations (CNV). SMAD4 deficit and DCC amplification are well known CNV in CRC, but they only appear in the high-ctDNA group. Similarly, CNV of AGO2, ACTG1, MYC, and BRD3 are only associated with the high-ctDNA group but not the low-ctDNA group. Conclusions: Our result indicates a strong correlation of fingerprint ctDNA level to both tumor mutational profile and prognosis. This suggests fingerprint ctDNA level may be used as an effective stratifying factor in CRC. At the same time, possibly a common molecular cause is driving persistent ctDNA production in CRC patients, including a large portion of them who have received surgical and adjuvant chemotherapy. Further perspective study, however, is needed to test whether fingerprint ctDNA can be a predictive biomarker.

MiST1: A phase IIa trial of rucaparib in patients harbouring BAP1/BRCA1 deficient relapsed malignant mesothelioma.

9057 Background: Malignant Mesothelioma (MM) remains an incurable cancer lacking effective treatments in the relapsed setting. Personalised therapy is still in its infancy. Homologous recombination (HR) deficiency associated with BRCA1 mutation has been shown to predict sensitivity to inhibition of poly-ADP ribose. In MM, BRCA1 associated protein 1 carboxy-terminal hydrolase (BAP1) is frequently mutated. It regulates both HR, and BRCA1 expression which is lost in 38% of mesotheliomas. Mesothelioma Stratified Therapy 1 (MiST1) was designed to test the hypothesis that BAP1/BRCA1 negative mesotheliomas would exhibit defective HR and exhibit sensitivity to PARP inhibition. Methods: MiST1 was a single centre, open label single arm phase IIa design with prospective molecular stratification; cytoplasmic/negative BAP1 or BRCA1 deficient MM was deemed eligible. Treatment was 600mg BD rucaparib (R) PO daily every 28 days for 6 cycles or until disease progression, unacceptable toxicity, withdrawal or death. Primary outcome was disease control rate at 12 weeks (DCR12w); secondary outcomes safety and toxicity profile, objective response rate (ORR) and DCR at 24 weeks (DCR24w). The null hypothesis states the true DCR12w is less than or equal to 25% and was tested against a one-sided alternative hypothesis that the DCR12w will be equal to or greater than 50%. Results: Between February 2019 and June 2019, 26 patients (pts) were eligible and consented to MiST1, median age 65.5 years, 85% are male and 15% female. Of these pts 15% had an ECOG performance status (PS) of 0 and 85% had an ECOG PS of 1. Molecular eligibility was 89% for BAP1 alone, 50% BRCA1 alone, and 39% BAP1+BRCA1. Primary tumour site was thoracic (96%) and subtype epithelioid (81%). DCR12w was 57.7% (95% CI, 36.9 - 76.7), DCR24w was 23.1% (95% CI, 9.0 – 43.7) and ORR was 11.5% (95%CI, 2.5 – 30.2). R was well tolerated with 9% (15/166) grade (G) 3/4 toxicities seen in 10 pts (38%), with no G5 toxicities. The most common adverse events were nausea occurring in 18 pts (69%), fatigue in 14 pts (54%), and decreased appetite in 10 pts (38%). Six cycles of R was received by 8 pts (30.8%). Dose reductions occurred in 9 pts (n = 8; 1 dose and n = 1; 2 doses). Dose delays occurred in 14 pts. Conclusions: MiST1 using the PARP inhibitor R met its primary endpoint of disease control rate at 12 weeks, showing promising efficacy with manageable toxicity. HR deficiency mutation signature enrichment is being investigated to refine the identification of responders to PARP inhibition. PARP inhibition warrants further investigation in MM. Clinical trial information: NCT03654833.

Molecular tumor diagnostics-current methods, examples of application and future

Molecular tumor diagnostics is rapidly evolving, driven by technological advances in high-throughput sequencing and bioinformatics, and paralleled by the emergence of novel therapeutic approaches, not least in the field of immune oncology. Molecular stratification of common tumor entities such as colorectal carcinoma and the discovery of paradigmatic molecular changes such as NTRK fusions illustrate how molecular pathological investigations can be performed under diagnostic or predictive conditions and can also provide prognostic information. Promising recent developments include "liquid biopsy" and tumor agnostic strategies. Afunctioning interdisciplinary cooperation between oncology, bioinformatics and molecular pathology is prerequisite for modern oncological diagnostics based on the current state-of-the-art knowledge and decisive for optimal care of oncological patients.