Search

Objective: Histology-independent (HI) technologies are authorised for advanced/metastatic cancer patients if they express a particular biomarker regardless of its position in the body. Although this represents an important advancement in cancer treatment, genomic testing to identify eligible individuals for HI technologies will require substantial investment and impact their cost-effectiveness. Estimating these costs is complicated by several issues which not only impact the overall cost of testing, but also on the distribution of testing costs across tumour types. Methods: Key issues that should be considered when evaluating the cost of genomic testing to identify those eligible for HI technology are discussed. These issues are explored in illustrative analyses where costs of genomic testing for NTRK fusions in England for recently approved HI technologies are estimated. Results: The prevalence of mutation, testing strategy adopted and current testing provision impact the cost of identifying eligible patients. The illustrative analysis estimated the cost of RNA-based NGS to identify one individual with an NTRK fusion ranged between £377 and £282,258. To improve cost-effectiveness, testing costs could be shared across multiple technologies. An estimated, additional ~4,000 patients would need to be treated with other HI therapies for testing in advanced/metastatic cancer patients to be cost-effective. Conclusions: The cost of testing to identify individuals eligible for HI technologies impact the drug’s cost-effectiveness. The cost of testing across tumour types varies owing to heterogeneity in the mutation’s prevalence and current testing provision. The cost-effectiveness of HI technologies may be improved if testing costs could be shared across multiple agents.

Additional file 2: Fig. S2. PARGi stabilises PAR chains in cell lines irrespective of PARGi sensitivity. (A) Quantification of PAR staining intensity in response to 48 h treatment with DMSO (Control), 1 μM PARGi, and co-treatment with 1 μM PARGi and 1 μM PARPi using single-cell immunofluorescence microscopy in 1 biological replicate (dot plots, 1000 cells shown per condition). (B) Representative immunofluorescence images of PAR staining from (A), in PARGi-resistant (COV318) and PARGi-sensitive (OVMANA) cell lines. Scale bar: 20 μm. (C) Quantification of PAR staining in response to PARGi or co-treatment with PARGi and PARPi, normalised to DMSO-treated cells (Control). Mean of ≥3 biological replicates. (D) Representative immunoblot showing PAR chain formation in response to 48 h treatment with DMSO (Control), or 1 μM PARGi. (E) PAR immunoblot in (D), without adjustment to show inter-line variation. (F) Mean quantification of PAR staining by immunoblotting (≥2 biological replicates). Statistics: 2-way ANOVA with Dunnett’s multiple comparisons test, selected comparisons were between drug treatments and DMSO control. Error bars represent SEM. * p

Additional file 9 Fig. S9. Low expression of mitotic or apoptotic gene sets does not identify OCMs sensitive to PARGi. (A) XY plot showing correlation between sum of Z scores for mitotic and DNA replication genes for each OCM (p

Additional file 4: Fig. S4. Comparison of NanoString with CCLE and Klijn datasets; DNA replication and mitotic gene expression correlation. (A) Box and whisker plots for DNA replication (red), mitotic (blue) and apoptotic (grey) gene Z scores in 10 cell line panel. Values on left side indicate Pearson R2 for CCLE or data from Klijn et al., compared with the NanoString analysis. (B) XY plot of the correlation between DNA replication and mitotic (red), and DNA replication and apoptotic (grey) gene expression. Z scores sums for epithelial cancer cell lines in Broad 2019 CCLE dataset [39]. Ten cell line panel indicated by black diamonds, note OV56 CCLE expression data does not reflect NanoString expression data (see text). Pearson R2 values indicated below the graph, all comparisons, p

Additional file 3: Fig. S3. PARGi sensitivity is accompanied by markers of replication stress and the DNA damage response. (A) Quantification of foci per nucleus after 48 h of 1 μM PARGi treatment (γH2AX and RPA1), or nuclear pKAP1 intensity after 72 h of 1 μM PARGi treatment using single-cell immunofluorescence microscopy in 1 biological replicate (dot plots, 1000 cells shown per condition). (B) Upper panel: Quantification of foci per nucleus after 48 h of 1 μM PARGi treatment (γH2AX and RPA1), or nuclear pKAP1 intensity after 72 h of 1 μM PARGi treatment; Lower panel: Results from upper panel normalised to DMSO-treated cells (Control) and represented as fold-change. Mean of ≥3 biological replicates. (C) Representative immunoblot for Chk1 and pChk1, following 48 h with 1 μM PARGi (Gi) or DMSO as a negative control (C), or for 2 h with 2 mM hydroxyurea (H) as a positive control. Tao1 serves as loading control. (D) Quantification of Li-COR pChk1 immunoblotting shown in Fig. 2C, mean of ≥3 biological replicates. Data are expressed as the increase resulting from treatment as a percentage of maximum response (achieved with hydroxyurea [H]) to correct for inter-line variation i.e. % PARGi (pChk1/Chk1)/H - % DMSO (pChk1/Chk1)/H. Statistics: 2-way ANOVA with Sidak multiple comparisons test (B, D), selected comparisons were between PARGi treated values and DMSO control within each cell line. Error bars represent SD (A), SEM (B, D). *p

Additional file 8: Fig. S8. PARGi suppresses mitotic entry in sensitive OCMs. (A) Cell fate profiling, showing cell behaviour over 120 h treatment with 1 μM PARGi or DMSO (Control). Each horizontal line represents a single cell, with the colours indicating cell behaviour. Following mitosis, one daughter cell was chosen at random to continue the analysis. No mitotic entry does not include death in interphase where mitosis does not take place. Abnormal mitosis includes cell division, abnormalities such as tripolar cell divisions, binuclear daughter cells and division of binuclear cells. Slippage is recorded where cells enter mitosis, then exit without division. Fusion was recorded where daughter cells appear to separate but subsequently join back together. Representative of 2 biological replicates. (B) Number of cells that fail to enter mitosis over 120 h (maximum n = 50) following 1 μM PARGi, 1 μM PARPi or DMSO (Control). Statistics: 2-way ANOVA with Dunnett’s multiple comparisons test, selected comparisons were between PARGi or PARPi treated cells versus DMSO control. Mean of 2 biological replicates. Error bars represent SEM. (C) Exemplar image of CENPF (indicating G2 cells) and Rad51 immunofluorescence staining of OCM.246 after 2 Gy X-ray ionising radiation followed by 24 h PARPi, ▼ indicates an HRP cell (positive for CENPF with > 5 Rad51 foci). Scale bar: 10 μm. Bar chart shows fold-change in % [Rad51 + CENPF+ cells/CENPF+ cells] in DMSO-treated cells versus cells treated with 2 Gy X-ray ionising radiation followed by 24 h 1 μM PARPi. Dotted line indicates a 2-fold change, above which cells are considered HRP (below it HRD).

Additional file 10: Supplementary Table 1. Clinical data, OCMs and primary tumour blocks. The table outlines the clinical data for the 25 new HGSOC OCMs screened for PARPi and PARGi sensitivity. Key: dx, diagnosis; FIGO, International Federation of Gynecology and Obstetrics; gBRCAm, germline BRCA1/2 mutation; VUS, variant of uncertain clinical significance; CTx, chemotherapy; Ref, platinum-refractory disease (tumour progression during or within 4 weeks of completing platinum therapy); Res, platinum-resistant (tumour progression between 4 weeks and 6 months from completing platinum therapy); Sens, platinum-sensitive (tumour progression ≥6 months from completing platinum therapy). CN, chemonaïve; OS, overall survival; mo, months; WT, wild-type; FR, frameshift; NS, nonsense; IF, immunofluorescence; IHC, immunohistochemistry; For IF and IHC: CK7 and PAX8, coloured box (present), white box (absent), S, strong; P, patchy; F, focal; W, weak. p53 is either mutant-type (strong/diffuse nuclear staining; darker coloured box), wild-type (lighter colour box) or absent nuclear staining (white box). NE, not evaluable (antibody failed); block unavailable (grey box). IB, immunoblotting; For IB: +(sm), band present but at lower than 53 kDa; +(S), strong band present; *At the time of the research biopsy; †Histologically re-classified from HGSOC to intermediate grade (grade 2/moderately differentiated) serous adenocarcinoma following tumour block analysis; ‡cell agar block only, histologically re-classified from HGSOC to suspicion of adenocarcinoma arising from the gynaecological tract. Clinical data for previously characterised OCMs are published [13], and not repeated here. Supplementary Table 2. BRCA1/2 variants detected in OCMs. The table outlines the BRCA1/2 variants detected in the panel of 7 OCMs screened for PARGi and PARPi sensitivity using colony survival and cell proliferation assays. Variants are described using Human Genome Variation Society (HGVS) nomenclature. Key: FR, frameshift; MS, missense; NS, nonsense; VAF, variant allele frequency; VUS, variant of uncertain clinical significance; WT, wild-type. All OCMs underwent both NGS and MLPA testing for BRCA1/2 variants. Supplementary Table 3. EC50 values for OCMs for the inhibitors/compounds tested. The table outlines the EC50 values for the inhibitors/compounds tested in the cell proliferation assay for the panel of 7 OCMs assessed. See Fig. 5C and D for Log EC50 values. Data are mean ± SEM from 3 biological replicates. PRISM could not accurately calculate EC50 for PARGi-resistant cells, therefore for resistant OCMs, the EC50 was approximated as 50 μM (half the maximal concentration tested) and described in the table as not calculable (NC).

Additional file 5: Fig. S5. Validation of patient-derived OCMs as bona fide models of HGSOC. (A) Representative images of severely atypical nuclei seen across OCMs. Scale bar: 10 μm. (B) Representative images of CK7 and PAX8, and p53 mutation-type (OCM.106: absent nuclear expression; OCM.195: strong/diffuse nuclear expression, involving > 80% tumour cell nuclei) by immunofluorescence staining. Scale bar: 20 μm. (C) Representative images of CK7, PAX and p53 mutation-type (OCM.92: absent nuclear expression; OCM.191: strong/diffuse nuclear expression, involving > 80% tumour cell nuclei) by immunohistochemistry staining from archival tumour blocks. Scale bar: 500 μm (× 10 magnification) and 100 μm (× 40 magnification). (D) Representative p53 immunoblot e.g. showing absent (OCM.86) and strong p53 bands (OCM.105). The control well represents stromal cells from the patient sample associated with OCM.237. Tao1 serves as loading control. (E) Representative somatic TP53 variants detected in OCMs. (F) Exemplar images of genome-wide chromosome copy-number profiles determined by single-cell whole-genome sequencing showing aneuploidies and rearranged chromosomes in tumour cells. Each row represents a single cell, with chromosomes plotted as columns and colours depicting copy-number state. See also Supplementary Table 1.

Additional file 6: Fig. S6. Living biobank screen demonstrated broad range of PARGi and PARPi sensitivity. (A) Colony formation following 96 h of treatment with 1 μM PARGi or 1 μM PARPi or DMSO (Control). (B) Quantification of ca from (A) following 1 μM PARGi (B) or 1 μM PARPi (C), normalised to DMSO-treated cells (Control) and represented as fold-change. BRCA status (germline, or, where available, OCM) indicated as follows: + = BRCA status of OCM; W=BRCA1/2 wild-type; B=BRCA1/2 mutation; V=BRCA1/2 variant of uncertain clinical significance; R = putative BRCA1/2 reversion; ▼=prior PARPi therapy. Single technical replicate. (D) Exemplar images of colony formation following continuous treatment with 1 μM PARGi or 1 μM PARPi or DMSO (Control). Representative images of 3 biological replicates. (E) Quantification of ca from (D) normalised to DMSO-treated cells (Control) and represented as fold-change. Mean of 3 biological replicates. Error bars represent SEM. See also Supplementary Table 2 and Fig. 5A, B. ***p

Additional file 7: Fig. S7. On-target inhibition of PARG in PARGi-sensitive OCMs. (A) Chemical structure of inactive small molecule analog of PARGi, PARGi-Me (PDD00031704). (B) Colony formation of PARGi-sensitive OCMs (109, 191 and 246) following continuous treatment with 1 μM PARGi or 1 μM PARGi-Me or DMSO (Control). Representative images of 3 biological replicates. (C) Upper panel – Proliferative Log EC50 values for PARGi and PARGi-ME for PARGi-sensitive OCMs. Mean of 3 biological replicates. Error bars represent SEM. Statistics: Unpaired t-test of PARGi versus PARGi-Me. Lower panel – Quantification of ca from (B) normalised to DMSO-treated cells (Control) and represented as fold-change. Mean of 3 biological replicates. Statistics: Unpaired t-test of PARGi versus PARGi-Me. Error bars represent SEM. ***p