Validation of Immunotherapy Response Score as Predictive of Pan-solid Tumor Anti-PD-1/PD-L1 Benefit

Immunotherapy response score (IRS) integrates tumor mutation burden (TMB) and quantitative expression biomarkers to predict anti-PD-1/PD-L1 [PD-(L)1] monotherapy benefit. Here, we evaluated IRS in additional cohorts. Patients from an observational trial (NCT03061305) treated with anti-PD-(L)1 monotherapy were included and assigned to IRS-High (-H) versus -Low (-L) groups. Associations with real-world progression-free survival (rwPFS) and overall survival (OS) were determined by Cox proportional hazards (CPH) modeling. Those with available PD-L1 IHC treated with anti-PD-(L)1 with or without chemotherapy were separately assessed. Patients treated with PD-(L)1 and/or chemotherapy (five relevant tumor types) were assigned to three IRS groups [IRS-L divided into IRS-Ultra-Low (-UL) and Intermediate-Low (-IL), and similarly assessed]. In the 352 patient anti-PD-(L)1 monotherapy validation cohort (31 tumor types), IRS-H versus IRS-L patients had significantly longer rwPFS and OS. IRS significantly improved CPH associations with rwPFS and OS beyond microsatellite instability (MSI)/TMB alone. In a 189 patient (10 tumor types) PD-L1 IHC comparison cohort, IRS, but not PD-L1 IHC nor TMB, was significantly associated with anti-PD-L1 rwPFS. In a 1,103-patient cohort (from five relevant tumor types), rwPFS did not significantly differ in IRS-UL patients treated with chemotherapy versus chemotherapy plus anti-PD-(L)1, nor in IRS-H patients treated with anti-PD-(L)1 versus anti-PD-(L)1 + chemotherapy. IRS associations were consistent across subgroups, including both Europeans and non-Europeans. These results confirm the utility of IRS utility for predicting pan-solid tumor PD-(L)1 monotherapy benefit beyond available biomarkers and demonstrate utility for informing on anti-PD-(L)1 and/or chemotherapy treatment.

Significance

This study confirms the utility of the integrative IRS biomarker for predicting anti-PD-L1/PD-1 benefit. IRS significantly improved upon currently available biomarkers, including PD-L1 IHC, TMB, and MSI status. Additional utility for informing on chemotherapy, anti-PD-L1/PD-1, and anti-PD-L1/PD-1 plus chemotherapy treatments decisions is shown.

Abstract 2171: A multivariate biomarker predicts sacituzumab govitecan response in solid tumors

Despite the recent clinical success of antibody drug conjugates (ADC) in oncology, predictive biomarkers are lacking, potentially limiting their impact. Herein, we evaluated the ability of candidate biomarkers alone and in combination to predict objective response rates observed in solid tumor patients treated with the TROP2-targeted ADC, sacituzumab govitecan (SG), as determined in the IMMU-12-01 basket trial. We leveraged available next generation sequencing (NGS)-based molecular profiling data from an independent advanced solid tumor cohort (n = 23,968) and developed a multivariate biomarker algorithm that produced biomarker positive rates correlating with the objective response rates (ORR) observed in IMMU-12-01. Candidate biomarkers evaluated included TROP2 gene expression, proliferation (by gene expression) and tumor cellularity. Notably, while TROP2 gene expression was highly correlated with protein expression across 45 tumor types (r = 0.93), TROP2 gene expression alone did not significantly predict ORR across 9 tumor types (r = 0.40, p = 0.29). In contrast, a biomarker algorithm combining TROP2 and proliferation by gene expression with tumor cellularity strongly predicted response both when using tumor type-specific biomarker rates in a discovery cohort (r = 0.83, p = 0.006) and in an independent validation cohort (r = 0.82, p = 0.007). These results indicate that the multivariate biomarker accounts for 67% of the variability observed in response rates and may thus identify patients likely to benefit from SG. Among tumor types with objective responses in IMMU-12-01, biomarker positive rates ranged from 9.9% in colorectal cancer to 57.4% in bladder cancer. Additional tumor types with biomarker positive rates >30% included cancers of the head and neck, cervix, salivary gland, skin (non-melanoma) and ovary, all with positive biomarker rates >30%. Interestingly, most tumor types had biomarker positive rates >5%, suggesting the potential for a tumor type-agnostic approach to patient selection. Considering SG and other ADC’s mechanism of action, a plausible model for response is that (1) higher target expression increases ADC drug delivery, (2) higher tumor cellularity increases ADC bystander effect and (3) higher tumor cell proliferation increases tumor cell death. In summary, we uncovered a novel biomarker algorithm capable of predicting SG response across solid tumors that may be generalizable to ADCs as a class, with the potential to further optimize use and maximize benefit.

Citation Format: Nickolay A. Khazanov, Laura E. Lamb, Daniel H. Hovelson, Kat Kwiatkowski, D. Bryan Johnson, Daniel R. Rhodes, Scott A. Tomlins. A multivariate biomarker predicts sacituzumab govitecan response in solid tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 2171.

Abstract 968: Evaluation of Her2 RNA expression as a potential predictive biomarker for anti-Her2 therapy

Trastuzumab deruxtecan (Enhertu) is effective in "HER2 Low" breast cancer, defined by 1+ or 2+ expression by immunohistochemistry (IHC). Interest has now turned toward defining a sub-population of IHC 0+ tumors that may have HER2 expression below the limit of IHC detection/quantification and may thus also be responsive. We previously validated a high dynamic range HER2 RNA expression assay run as part of our comprehensive genomic profiling test, StrataNGS.

Herein, we evaluated the HER2 RNA expression data together with copy number and clinical outcome data from the Strata Clinical Molecular Database (SCMD) in advanced breast cancer (n = 3,063) and other advanced solid tumors (n = 26,715). As expected, HER2 gene expression was significantly higher in tumors with DNA amplification (>=6 copies; median: 13.9 vs. 10.0 in log2 units; p < 1e-100). Despite similar copy number levels in amplified breast vs. other cancers (median: 21.8 vs. 19.8 copies), HER2 expression levels were ~2-fold higher (median: 14.5 vs. 13.5; p = 1.3e-10). Similarly, HER2 expression levels were higher in non-amplified breast vs. other cancers (median: 10.7 vs. 9.9; p<1e-100), suggesting that DNA amplification and cell lineage affect HER2 expression. Using our previously validated HER2 threshold, among 75 eligible SCMD breast cancer patients treated with 1st or 2nd line systemic trastuzumab or pertuzumab containing therapy, HER2 RNA High patients (n=46, 59%) had significantly longer time to next therapy (TTNT) compared to HER2 RNA Not High patients (median TTNT 26.9 vs. 5.6 months, adjusted hazard ratio 0.31, p=0.005 when adjusted for 1st vs. 2nd line, pertuzumab inclusion, and inclusion of chemotherapy or hormonal therapy).

In patients with available IHC data (n = 388), HER2 RNA expression trended with IHC across the 0-3+ range, however, while 3+ tumors had distinctly high RNA expression (median: 14.4), 0-2+ tumors had lower expression with overlapping distributions (median: 10.5, 10.9, 11.5, respectively), suggesting that 0-2+ tumors do not represent distinct biological groups, but rather a continuum of low expression. We defined a HER2 RNA Low threshold (>10.6), corresponding to the top 75% of IHC 1-2+ breast cancers. Importantly, at this threshold, nearly half (44.1%) of 0+ breast cancers were also classified as HER2 RNA Low. Additionally, 25.8% of all non-breast solid tumors were classified as HER2 RNA Low. Given that HER2 RNA High predicted benefit from 1st generation anti-HER2 therapies, future studies should consider HER2 RNA Low as an alternative biomarker to Her2 IHC Low, with the opportunity to further expand trastuzumab deruxtecan use into the IHC 0+ breast cancer population and potentially to additional solid tumors.

Citation Format: Laura E. Lamb, Nickolay A. Khazanov, Daniel H. Hovelson, Kat Kwiatkowski, D. Bryan Johnson, Daniel R. Rhodes, Scott A. Tomlins. Evaluation of Her2 RNA expression as a potential predictive biomarker for anti-Her2 therapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 968.

Analytical and Potential Clinical Performance of Oncomine Myeloid Research Assay for Myeloid Neoplasms

INTRODUCTION

Next-generation sequencing (NGS) panels have recently been introduced to efficiently detect genetic variations in hematologic malignancies.

OBJECTIVES

Our aim was to evaluate the performance of the commercialized Oncomine™ myeloid research assay (OMA) for myeloid neoplasms.

METHODS

Certified reference materials and clinical research samples were used, including 60 genomic DNA and 56 RNA samples. NGS was performed using OMA, which enables the interrogation of 40 target genes, 29 gene fusions, and five expression target genes with five expression control genes by the Ion S5 XL Sequencer. The analyzed data were compared with clinical data using karyotyping, reverse transcription polymerase chain reaction (PCR), fluorescence in situ hybridization, Sanger sequencing, customized NGS panel, and fragment analysis.

RESULTS

All targets of reference materials were detected except three (two ASXL1 and one CEBPA) mutations, which we had not expected OMA to detect. In clinical search samples, OMA satisfactorily identified DNA variants, including 90 single nucleotide variants (SNVs), 48 small insertions and deletions (indels), and eight FLT3 internal tandem duplications (ITDs) (Kappa agreement 0.938). The variant allele frequencies of SNVs and indels measured by OMA correlated well with clinical data, whereas those of FLT3-ITDs were significantly lower than with fragment analysis (P = 0.008). Together, OMA showed strong ability to identify RNA gene fusions (Kappa agreement 0.961), except one RUNX1-MECOM. The MECOM gene was highly expressed in all five samples with MECOM-associated rearrangements, including inv(3), t(3;3), and t(3;21).

CONCLUSION

OMA revealed excellent analytical and potential clinical performance and could be a good replacement for conventional molecular tests.

Exploring the composition of protein-ligand binding sites on a large scale

The residue composition of a ligand binding site determines the interactions available for diffusion-mediated ligand binding, and understanding general composition of these sites is of great importance if we are to gain insight into the functional diversity of the proteome. Many structure-based drug design methods utilize such heuristic information for improving prediction or characterization of ligand-binding sites in proteins of unknown function. The Binding MOAD database if one of the largest curated sets of protein-ligand complexes, and provides a source of diverse, high-quality data for establishing general trends of residue composition from currently available protein structures. We present an analysis of 3,295 non-redundant proteins with 9,114 non-redundant binding sites to identify residues over-represented in binding regions versus the rest of the protein surface. The Binding MOAD database delineates biologically-relevant “valid” ligands from “invalid” small-molecule ligands bound to the protein. Invalids are present in the crystallization medium and serve no known biological function. Contacts are found to differ between these classes of ligands, indicating that residue composition of biologically relevant binding sites is distinct not only from the rest of the protein surface, but also from surface regions capable of opportunistic binding of non-functional small molecules. To confirm these trends, we perform a rigorous analysis of the variation of residue propensity with respect to the size of the dataset and the content bias inherent in structure sets obtained from a large protein structure database. The optimal size of the dataset for establishing general trends of residue propensities, as well as strategies for assessing the significance of such trends, are suggested for future studies of binding-site composition.

Describing the general structure of protein binding sites is fundamentally important for guiding drug design and better understanding structure-function relationships. Here, we analyze small molecules bound to proteins within our large database, Binding MOAD (Mother of All Databases, pronounced like “mode” as a pun referring to ligand-binding modes). We focus on different contacts across the residues in the binding sites, and we normalize the data relative to the protein's entire surface. A key feature of this study is the use of a “control” where we compare real, functional binding sites to the random contacts seen for crystallographic additives against the protein surface. Controls are required in experimental biology, but they are ill-defined in many computational approaches. This allows us to describe how true binding sites are unique on the protein surface and distinct from random patches that attract common, small molecules.

Abstract C256: Expanded clinical opportunities for crizotinib from an analysis of over 5,000 cancer patient exomes

Patients with chromosomal rearrangements resulting in fusion proteins are amongst the most responsive to targeted therapy. For example, targeting of the BCR-ABL fusion in chronic myelogenous leukemia (CML) with imatinib and the EML4-ALK fusion in non-small cell lung cancer (NSCLC) with crizotinib has led to dramatic patient responses in these diseases. While crizotinib is approved for use in EML4-ALK positive NSCLC through its inhibition of ALK, the drug also inhibits ROS1, MST1R (RON), MET, and more recently has been shown to inhibit the ALK homolog, LTK. To gain a more comprehensive understanding of the full therapeutic potential of crizotinib, we undertook a genomic survey of ALK, LTK, ROS1, MET and MST1R across thousands of patients subjected to full exome sequencing including patients from The Cancer Genome Atlas (TCGA) and the International Cancer Genome Consortium (ICGC), as well as tens thousands of patients from Oncomine® databases. We confirmed the presence of EML4-ALK fusions in both lung and colorectal cancer (CRC), identified a PRKAR1A-ALK fusion in CRC, and found evidence of novel recurrent ALK fusions in kidney papillary renal cell carcinoma and thyroid gland carcinoma. ALK hotspot mutations and focal amplifications were confined to neuroblastoma, as previously described. We also report the first instance of an LTK fusion, identified in thyroid gland carcinoma. LTK amplifications were also observed in 1.4% of gastric cancers and rarely in medulloblastoma and breast cancer. LTK was prominently over-expressed in leukemia, and in an analysis of over 4,000 PML-RARA fusion positive leukemia patients, LTK was amongst the most significantly over-expressed genes. In addition to confirming previously published ROS1 fusions, our survey of ROS1 identified rare novel fusions in NSCLC and glioblastoma. High-level MET amplifications were observed in 1-5% of papillary renal cell carcinoma, the intestinal subtype of gastric adenocarcinoma, oligodendroglioma, glioblastoma and lung adenocarcinoma. Hotspot mutations in MET were frequently observed in head and neck squamous cell carcinoma (HNSCC) (11%), and observed in a third of metastatic HNSCC samples. Additional hotspot mutations were also observed in lung adenocarcinoma (2%) and small cell lung cancer (2%). Aberrations in MST1R were rare. These results leverage all available genomic profiling data to provide a broadened scope of therapeutic opportunity for inhibitors like crizotinib. With the growing availability of next-generation sequencing data and analyses, such surveys can support hypothesis-driven development of targeted therapies and help expand opportunities for clinical stage therapies.

Citation Information: Mol Cancer Ther 2013;12(11 Suppl):C256.

Citation Format: Sean Eddy, Mark Tomilo, Mary Ellen Urick, Nickolay A. Khazanov, Paul Williams, Armand Bankhead, Dinesh Cyanam, Supra Gajjala, Peter Wyngaard, Emma Bowden, Dan R. Rhodes. Expanded clinical opportunities for crizotinib from an analysis of over 5,000 cancer patient exomes. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2013 Oct 19-23; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2013;12(11 Suppl):Abstract nr C256.

Abstract 3173: High-throughput, systematic analysis of paired-end next-generation sequencing data to characterize the gene fusion landscape in cancer

Gene fusions encode oncogenic drivers in hematological and solid tumors and are often associated with dramatic clinical responses with the appropriate targeted agents. In principle, massively parallel paired-end sequencing can identify structural rearrangements in tumor genomes and transcriptomes. However, computational methods to identify gene fusions are varied, still evolving and largely trained on cell line data. We sought to develop systematic methods to characterize known oncogenic gene fusions and to discover novel gene fusions in cancer. RNASeq data for approximately 3,400 clinical cases from 16 cancer types was obtained from the Cancer Genomics Hub (CGHub) of The Cancer Genome Atlas (TCGA). We surveyed the performance of several gene fusion callers and chose two (deFuse and TopHat) for further method development. An analysis pipeline was developed and executed in parallel on a high-performance computing cluster. Filtering and annotation was conducted on the aggregated data as a post-processing step, to enable exploratory analyses of various filters. We optimized filtering approaches on datasets that included known standards (e.g., TMPRSS2-ERG in prostate adenocarcinoma, PML-RARA in acute myeloid leukemia, etc.) to enrich for these and other gene fusions with correct 5’-3’ orientation while excluding cases with ambiguous breakpoints and spanning reads, alignment errors, and read-through transcripts from adjacent genes. Predicted fusions were summarized based on the occurrence of unique genes participating in fusions with multiple partners and of unique gene pairs, each within specific diseases. Elevated expression was observed after the predicted breakpoint of the 3’ gene in cases positive for predicted fusions, and added important confirmatory evidence. Thus, we characterized the incidence and distribution of several known oncogenic gene fusions including EML4-ALK and CCDC6-RET while expanding the number of gene partners identified in combination with oncogenes such as ROS1. In addition to characterizing the incidence and distribution of 31 known gene fusions, we nominated over 100 novel gene fusion pairs. One example of a novel gene fusion susceptible to available targeted therapy was FGFR3-TACC3 in 4% of bladder cancer, 2% of squamous cell lung carcinoma, and 1% each of glioblastoma and head and neck squamous cell carcinoma. Computational methods are now poised to complement biochemical approaches in the definition of the gene fusion landscape in cancer.

Citation Format: Seth E. Sadis, Nickolay A. Khazanov, Armand R. Bankhead, Dinesh Cyanam, Paul D. Williams, Sean F. Eddy, Peter J. Wyngaard, Daniel R. Rhodes. High-throughput, systematic analysis of paired-end next-generation sequencing data to characterize the gene fusion landscape in cancer. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3173. doi:10.1158/1538-7445.AM2013-3173

Overcoming sequence misalignments with weighted structural superposition

An appropriate structural superposition identifies similarities and differences between homologous proteins that are not evident from sequence alignments alone. We have coupled our Gaussian-weighted RMSD (wRMSD) tool with a sequence aligner and seed extension (SE) algorithm to create a robust technique for overlaying structures and aligning sequences of homologous proteins (HwRMSD). HwRMSD overcomes errors in the initial sequence alignment that would normally propagate into a standard RMSD overlay. SE can generate a corrected sequence alignment from the improved structural superposition obtained by wRMSD. HwRMSD's robust performance and its superiority over standard RMSD are demonstrated over a range of homologous proteins. Its better overlay results in corrected sequence alignments with good agreement to HOMSTRAD. Finally, HwRMSD is compared to established structural alignment methods: FATCAT, secondary-structure matching, combinatorial extension, and Dalilite. Most methods are comparable at placing residue pairs within 2 Å, but HwRMSD places many more residue pairs within 1 Å, providing a clear advantage. Such high accuracy is essential in drug design, where small distances can have a large impact on computational predictions. This level of accuracy is also needed to correct sequence alignments in an automated fashion, especially for omics-scale analysis. HwRMSD can align homologs with low-sequence identity and large conformational differences, cases where both sequence-based and structural-based methods may fail. The HwRMSD pipeline overcomes the dependency of structural overlays on initial sequence pairing and removes the need to determine the best sequence-alignment method, substitution matrix, and gap parameters for each unique pair of homologs.

CSAR benchmark exercise of 2010: selection of the protein-ligand complexes

A major goal in drug design is the improvement of computational methods for docking and scoring. The Community Structure Activity Resource (CSAR) aims to collect available data from industry and academia which may be used for this purpose (www.csardock.org). Also, CSAR is charged with organizing community-wide exercises based on the collected data. The first of these exercises was aimed to gauge the overall state of docking and scoring, using a large and diverse data set of protein–ligand complexes. Participants were asked to calculate the affinity of the complexes as provided and then recalculate with changes which may improve their specific method. This first data set was selected from existing PDB entries which had binding data (K_d or K_i) in Binding MOAD, augmented with entries from PDBbind. The final data set contains 343 diverse protein–ligand complexes and spans 14 pK_d. Sixteen proteins have three or more complexes in the data set, from which a user could start an inspection of congeneric series. Inherent experimental error limits the possible correlation between scores and measured affinity; R² is limited to ∼0.9 when fitting to the data set without over parametrizing. R² is limited to ∼0.8 when scoring the data set with a method trained on outside data. The details of how the data set was initially selected, and the process by which it matured to better fit the needs of the community are presented. Many groups generously participated in improving the data set, and this underscores the value of a supportive, collaborative effort in moving our field forward.

Impact of sex, age, race, ethnicity and aspirin use on bleeding symptoms in healthy adults

BACKGROUND

Comparing a patient's bleeding symptoms with those of healthy individuals is an important component of the diagnosis of bleeding disorders, but little is known about whether bleeding symptoms in healthy individuals vary by sex, race, ethnicity, age, or aspirin use. OBJECTIVES, PATIENTS/METHODS: We developed a comprehensive, ontology-backed, Web-based questionnaire to collect bleeding histories from 500 healthy adults. The mean age was 43 years (range 19-86 years), 63% were female, 19% were Hispanic, 37% were African-American, 43% were Caucasian, 8% were Asian, and 4% were multiracial.

RESULTS

18 of the 36 symptoms captured occurred with < 5% frequency, and 26% of participants reported no bleeding symptoms (range 0-19 symptoms). Differences in sex, race, ethnicity, aspirin use and age accounted for only 6-13% of the variability in symptoms. Although men reported fewer symptoms than women (median 1 vs. 2, P < 0.01), there was no difference when sex-specific questions were excluded (median 1 for both men and women, P = 0.50). However, women reported more easy bruising (24% vs. 7%, P < 0.01) and venipuncture-related bruising (10% vs. 3%, P = 0.02). The number of symptoms did not vary by race or age, but epistaxis was reported more frequently by Caucasians than by African-Americans (29% vs. 18%, P = 0.02), and epistaxis frequency decreased with age (odds ratio 0.97 per year, P < 0.01). Paradoxically, infrequent aspirin users reported more bruising and heavy menses than frequent users (21% vs. 8%, P = 0.01, and 56% vs. 38%, P = 0.03, respectively).

CONCLUSIONS

Our findings provide a contemporaneous and comprehensive description of bleeding symptoms in a diverse group of healthy individuals. Our Web-based system is freely available to other investigators.

Differences between high- and low-affinity complexes of enzymes and nonenzymes

Physical differences in small molecule binding between enzymes and nonenzymes were found through mining the protein-ligand database, Binding MOAD (Mother of All Databases). The data suggest that divergent approaches may be more productive for improving the affinity of ligands for the two classes of proteins. High-affinity ligands of enzymes are much larger than those with low affinity, indicating that the addition of complementary functional groups is likely to improve the affinity of an enzyme inhibitor. However, this process may not be as fruitful for ligands of nonenzymes. High- and low-affinity ligands of nonenzymes are nearly the same size, so modest modifications and isosteric replacement might be most productive. The inherent differences between enzymes and nonenzymes have significant ramifications for scoring functions and structure-based drug design. In particular, nonenzymes were found to have greater ligand efficiencies than enzymes. Ligand efficiencies are often used to indicate druggability of a target, and this finding supports the feasibility of nonenzymes as drug targets. The differences in ligand efficiencies do not appear to come from the ligands; instead, the pockets yield different amino acid compositions despite very similar distributions of amino acids in the overall protein sequences.

Binding MOAD, a high-quality protein-ligand database

Binding MOAD (Mother of All Databases) is a database of 9836 protein-ligand crystal structures. All biologically relevant ligands are annotated, and experimental binding-affinity data is reported when available. Binding MOAD has almost doubled in size since it was originally introduced in 2004, demonstrating steady growth with each annual update. Several technologies, such as natural language processing, help drive this constant expansion. Along with increasing data, Binding MOAD has improved usability. The website now showcases a faster, more featured viewer to examine the protein-ligand structures. Ligands have additional chemical data, allowing for cheminformatics mining. Lastly, logins are no longer necessary, and Binding MOAD is freely available to all at http://www.BindingMOAD.org.