A review of the cost-effectiveness of genetic testing for germline variants in familial cancer

Abstract

Background

Targeted germline testing is recommended for those with or at risk of breast, ovarian, or colorectal cancer. The affordability of genetic sequencing has improved over the past decade, therefore the cost-effectiveness of testing for these cancers is worthy of reassessment.

Objective

To systematically review economic evaluations on cost-effectiveness of germline testing in breast, ovarian, or colorectal cancer.

Methods

A search of PubMed and Embase databases for cost-effectiveness studies on germline testing in breast, ovarian, or colorectal cancer, published between 1999 and May 2022. Synthesis of methodology, cost-effectiveness, and reporting (CHEERS checklist) was performed.

Results

The incremental cost-effectiveness ratios (ICERs; in 2021-adjusted US$) for germline testing versus the standard care option in hereditary breast or ovarian cancer (HBOC) across target settings were as follows: (1) population-wide testing: 344–2.5 million/QALY; (2) women with high-risk: dominant = 78,118/QALY, 8,337–59,708/LYG; (3) existing breast or ovarian cancer: 3,012–72,566/QALY, 39,835/LYG; and (4) metastatic breast cancer: 158,630/QALY. Likewise, ICERs of germline testing for colorectal cancer across settings were: (1) population-wide testing: 132,200/QALY, 1.1 million/LYG; (2) people with high-risk: 32,322–76,750/QALY, dominant = 353/LYG; and (3) patients with existing colorectal cancer: dominant = 54,122/QALY, 98,790–6.3 million/LYG. Key areas of underreporting were the inclusion of a health economic analysis plan (100% of HBOC and colorectal studies), engagement of patients and stakeholders (95.4% of HBOC, 100% of colorectal studies) and measurement of outcomes (18.2% HBOC, 38.9% of colorectal studies).

Conclusion

Germline testing for HBOC was likely to be cost-effective across most settings, except when used as a co-dependent technology with the PARP inhibitor, olaparib in metastatic breast cancer. In colorectal cancer studies, testing was cost-effective in those with high-risk, but inconclusive in other settings. Cost-effectiveness was sensitive to the prevalence of tested variants, cost of testing, uptake, and benefits of prophylactic measures. Policy advice on germline testing should emphasize the importance of these factors in their recommendations.

PLAIN LANGUAGE SUMMARY

Breast, ovarian, prostate, and colorectal cancers are among the top causes of cancer related deaths. A substantial proportion of people with these cancers have inherited mutations. The identification of these gene abnormalities could provide people with opportunities to utilize preventive risk reduction surgeries or undertake frequent routine testing for these cancers. However, genetic testing requires healthcare resources and money. Previous reviews on the cost-effectiveness of genetic testing in familial cancers have concluded that targeted screening i.e., selective assessment of people at high-risk could justify the costs of testing. Our evaluation of economic studies in breast and ovarian cancer, however, suggests that genetic testing is cost-effective across a wide variety of situations starting from the screening of all healthy women above 30 years to the testing of women with existing breast or ovarian cancer. Testing in metastatic breast cancer to inform treatment with Olaparib, a drug known to selectively improve survival in people with genetic mutations, was the sole exception where testing was not cost-effective. Contrary to findings for breast or ovarian cancer, testing for colorectal cancer was cost-effective in people with high-risk i.e., family history but inconclusive in other situations. Evidence on the cost-effectiveness of testing in prostate cancer is lacking and as a result we were not able to provide advice in this cancer group.

Keywords:

• Germline testing
• breast cancer
• ovarian cancer
• colorectal cancer
• prostate cancer
• cost-effectiveness

Systematic review registration:

• CRD42020161967

JEL CLASSIFICATION CODES:

• I00
• I
• A10
• A1
• A

Introduction

Genetic testing is recommended for the risk assessment and management of hereditary cancer, including breast, ovarian, and colorectal cancers^1–3. Between 10% and 20% of these cancers have been linked with pathogenic germline variants (breast: 10%; ovarian: 20%; colorectal: 10%)^4–7. The underlying aetiological mechanisms for these cancers comprise a failure of homologous recombination repair (HRR) or DNA mismatch repair (dMMR). Carriers of pathogenic variants (BRCA and lynch syndrome variants) tend to present with cancer at an early age^8,9 and have an aggressive disease profile with high mortality¹⁰. Identifying patients at increased hereditary risk may provide opportunities for early management, such as intensive surveillance with routine screening (e.g. colonoscopy) in colorectal cancer¹¹ or pro-active treatments like prophylactic risk-reduction surgeries in women predisposed to breast or ovarian cancer¹².

Previous reviews^13–15 on the economics of genetic testing in breast, ovarian, and colorectal cancer indicate that testing is likely to be cost-effective when targeted towards populations with a higher-than-average risk of being a carrier or if multigene assays were utilized in the analysis. Findings from the review by Koldekoff et al.¹⁴ on breast and ovarian cancer studies, however, may be limited due to the authors criteria to selectively assess studies that examined germline testing in healthy women with high-risk and in women with existing breast and ovarian cancer with a cascade testing component. Studies on population-wide testing of women within the general population, testing of racial/ethnic groups with high-risk (Ashkenazi Jewish women) and studies that failed to include cascade testing were excluded. On the contrary, DiMarco et al.¹³ examined the cost-effectiveness of genetic testing for colorectal cancer over a wider range of populations and concluded that universal or targeted testing may be cost-effective when performed before 70 years of age.

There have been several additional studies on the cost-effectiveness of germline testing since the previous reviews^16–26. Some of them assessed population-wide testing^17,20, while others evaluated the cost-utility of germline testing guided treatments in cancer patients with mutations²⁶. Next-generation sequencing (NGS) has also become affordable²⁷. Therefore, it would be interesting to see if the reduction in testing costs expands the utility of germline testing to populations where testing was previously known to be beneficial but cost-prohibitive. In this context, the aim of the current systematic review was to undertake a contemporary and comprehensive assessment of economic evaluations that studied the cost-effectiveness of germline testing for breast, ovarian, prostate, and colorectal cancer.

Materials and methods

Inclusion criteria and search strategy

A protocol (CRD42020161967) for this systematic review was registered at the International Prospective Register of Systematic Reviews (PROSPERO). A comprehensive search of electronic databases at Medline (PubMed) and Embase (Elsevier) was performed in accordance with the Preferred Reporting System for Systematic Reviews and Meta-Analysis (PRISMA) guidelines²⁸. A checklist of the PRISMA items has been provided in Supplementary Table S1. We examined full economic evaluations (studies that had both costs and effectiveness measures) that assessed germline testing for breast, ovarian, prostate, and colorectal cancer in the databases between 1 January 1999 and 6 May 2022. The following search terms were used: genetic testing, germline testing, breast cancer, prostate cancer, colorectal cancer, lynch syndrome, cost-effectiveness analysis, cost-benefit analysis, economics, Markov, and decision support techniques. Details about the search strategy with MESH terms, PICO, inclusion, and exclusion criteria have been provided in Supplementary Table S2. To be brief, studies were included if they met the following inclusion criteria:

• Patients: (1) All adults in the general population, at risk of breast, ovarian, prostate, or colorectal cancer, or (2) patients with a pre-established diagnosis of breast, ovarian, prostate, or colorectal cancer.

• Intervention: Germline testing followed by personalized management including (1) intensive surveillance (colonoscopy, faecal immunohistochemistry, mammograms, etc.), (2) risk-reduction surgeries, or (3) specific pharmacotherapy, selective for cancer pathways (PARP inhibitors, EGFR inhibitors, etc.).

• Comparator: standard clinical care (1) without germline testing, (2) selective testing of individuals at high-risk, or (3) reflex germline testing after immunohistochemistry

• Outcome: costs per life year gained (LYG) or costs per quality-adjusted life year (QALY).

Data abstraction, synthesis, and quality assessment

Parameters relevant to germline testing and cost-effectiveness, including details about the country/setting, type of cancer, population for germline testing, referent population, pathogenic variants assessed, probability of being variant positive, cost of testing, cascade testing, type of prophylaxis in test population versus referent, compliance rates of prophylaxis or frequency of intensive surveillance, model used in the analysis, study perspective, discounting, incremental cost-effectiveness ratio (ICER), willingness-to-pay (WTP) threshold, and factors affecting cost-effectiveness (sensitivity analysis) were extracted from each study. We grouped results for hereditary breast and ovarian cancer (HBOC) together for reasons of sample size, on aetiological grounds (association with defects in homologous recombination repair mechanics)²⁹ and similarities in prophylactic measures (risk-reduction mastectomy, i.e. RRM; risk-reduction salpingo-oophorectomy, i.e. RRSO; or both). Studies in colorectal cancer evaluated several diagnostic strategies including no testing, reflex testing, as well as direct germline testing. Reflex testing begins with tumour-based screening that requires a tissue biopsy sample and is followed by germline testing only after a positive IHC test for MLH1 or BRAF. On the contrary, direct germline testing is less invasive and is typically performed on peripheral blood or saliva samples. Therefore, when reporting the ICERs for colorectal cancer studies we have strived to compare germline testing versus no testing, as the two strategies are more in line with the criteria for the review. Reflex testing was used as the referent, only in studies where it was the sole comparator for germline testing. All reported costs have been standardized from values reported in the original article to 2021 US$ using the CCEMG-EPPI-Centre Cost Converter³⁰. A meta-analysis was not possible as there was substantial heterogeneity among included studies in the target population for testing.

Reporting quality of the studies was assessed by two authors (ST and BH) using the recently updated Consolidated Health Economic Evaluation Reporting Standards (CHEERS) checklist³¹. These guidelines received a major overhaul in 2022 and, compared to the previous iteration, include new items on patient engagement and the use of a health economic analysis plan (HEAP). The goal of the CHEERS assessment was to report the essential information needed for interpretation of health economic evaluations, check for reliability of study findings, efforts made for patient or community engagement, and offer clarity to readers by providing the location of the 28 assessed items within each article.

Results

Search strategy and the final sample of studies

An initial search of the databases using a pre-defined set of keywords (Supplementary Table S2) resulted in the identification of 1,201 potential articles (Figure 1). Removal of duplicates and title screening brought this sample down to 126 articles. Further screening of abstracts and full texts led to the exclusion of an additional 93 articles (46 conference abstracts or letters, 19 not germline testing, 22 not full economic evaluations, two studies based on other cancers, one study not in English language, three studies using non-relevant comparators, one study in non-adult population). At this point of screening, we manually added seven articles that were not identified through the database search, five of which were articles published within the last year and two other studies that were present in a previous review¹⁵. The final set of articles included in the review was 40 (22 HBOC, 18 colorectal cancer). There were no studies on the cost-effectiveness of germline testing in prostate cancer. Tables 1 and 2 present a detailed overview of the characteristics of all the articles included in the review.

Figure 1. PRISMA flow chart.

Table 1. Characteristics of economic studies on germline testing for breast and ovarian cancer.

#	Study with publication year (Setting)	Target population for germline testing (Referent strategy)	Testing strategy, genes assessed, variant prob%; cost of test	Cascade testing	Prophylaxis (uptake rate)	Model (perspective), discounting %	ICER*; US$/QALY or LYG (WTP threshold)	Sensitivity analysis, i.e. factors affecting cost-effectiveness
1	Sun et al. (2022)^25 (China)	Women with unselected breast cancer (No testing)	NGS of BRCA 1, BRCA 2 & PALB2, 6.0%, US$ 367	Yes, first & second-degree relatives	CPM (54%), RRSO (57%)	Micro-simulation (Payer & Societal), 3%	Payer: 7,572/QALY, Societal: 4,696/QALY (US$ 10,262; 1x GDP)	None
2	Michaan et al. (2021)^20 (US)	Population-wide of healthy women ≥ 30 years (No testing)	TVA of BRCA 1 & 2 founder mutations; 2.4%; US$ 154	No	RRM (28%), RRSO (65%)	Markov (Payer), 3%	21,278/QALY (US$ 100,000)	Cost of testing, probability of variant positive, prophylaxis uptake, discount rate
3	Manchanda et al. (2020)^32 (UK, USA, Netherlands, China, Brazil, India)	Population-wide of healthy women ≥ 30 years (Testing in high-risk)	NGS of BRCA 1 & 2; 0.7%; $200	No	RRM (47%), RRSO (55%)	Markov (Payer & Societal), 3%	Payer: 18,293/QALY, Societal: dominant (US$ 100,000)	Cost of testing, especially in developing countries
4	Hurry et al. (2020)^33 (Canada)	Women with unselected breast or epithelial ovarian cancer (No testing)	NGS of BRCA 1 & 2; breast: 4.7%, ovarian: 12.5%; US$ 478	Yes, first-degree relatives	RRM (88%), RRM+RRSO (84%)	Microsimulation (Payer), 1.5%	11,665/QALY (US$ 100,000)	Uptake of prophylaxis
5	Guzauskas et al. (2020)^34 (US)	Population-wide of healthy women ≥ 30 years (Testing in high-risk)	NGS of BRCA1, BRCA2, ATM, CHEK2, MSH6, PALB2, RAD51C & TP53; 49.5%; US$200	Yes, first-degree relatives	RRM (46%), RRSO (86%)	Markov (Payer), 3%	91,392/QALY (US$ 100,000)	Cost of testing, prophylaxis uptake, benefits of prophylaxis, cascade testing
6	Sun et al. (2019)^35 (US, UK, Australia)	Women with unselected breast cancer (Testing in high-risk)	NGS of BRCA1, BRCA2 & PALB2; 5.5%; US$330	Yes, first & second-degree relatives	CPM (54%), RRSO (57%)	Microsimulation (Payer & Societal), 3.5%	Payer: 72,566/QALY, Societal: 68,098/QALY (US$ 100,000)	Cost of testing, uptake of prophylaxis
7	Saito et al. (2019)^26 (Japan)	Women with metastatic breast cancer with HER-2 or triple negative status (No testing)	NGS of BRCA 1 & 2; 11.3%; US$2,058	No	PARP Inhibitor vs. standard chemo	Markov (Payer), 2%	158,630/QALY (US$ 129,695, 3× GDP)	Survival rate with PARP therapy
8	Moya-Alarcon et al. (2019)^36 (Spain)	Women with non-mucinous high-grade epithelial ovarian cancer (No testing)	TVA of BRCA 1 & 2; 8.3%; US$716	Yes, first-degree relatives	RRM (20%), RRSO (20–65%)	Microsimulation (Payer), 3%	52,290/QALY (US$ 82,683 or €50,000)	None
9	Tuffaha et al. (2018)^37 (Australia)	Women with breast cancer and with high-risk of mutations (No testing)	NGS of BRCA 1 & 2; 15.0%; US$935	Yes, first & second-degree relatives	RRM (30%), RRSO (54%)	Markov (Payer), 5%	14,730/QALY (US$ 34,770, AU$ 50,000)	None
10	Patel et al. (2018)^38 (UK, US)	Population wide of healthy Sephardi Jewish women ≥ 30 years (Testing in high-risk)	TVA of BRCA 1 founder mutation; 0.7%; US$300	No	RRM (60%), RRSO (66%)	Markov (Payer), 3.5%	344/QALY (US$ 100,000)	None
11	Manchanda et al. (2018)^39 (UK, US)	Population-wide of healthy women ≥ 30 years (Testing in high-risk)	NGS of BRCA1, BRCA2, RAD51C, RAD51D, BRIP1, PALB2; 0.7%; US$330	No	RRM (47%), RRSO (55%)	Decision-tree (Payer), 3.5%	61,846/QALY (US$ 100,000)	None
12	Lim et al. (2018)^40 (Malaysia)	Women with early-stage breast cancer (No testing)	NGS of BRCA 1 & 2; 13.7%; US$451	No	RRM (21%), RRSO (48%)	Decision-tree (Payer), 3%	3,012/QALY (US$ 9,500, 1× GDP)	Cost of testing, probability of variant positive, prophylaxis uptake, benefit of prophylaxis, discount rate
13	Manchanda et al. (2017)^41 (UK, US)	Population wide of healthy women ≥ 30 years with Ashkenazi Jewish grandparents (Testing only women with high-risk)	TVA of BRCA 1 & 2 founder mutations; 2.5%; US$300	No	RRM (52%), RRSO (55%)	Decision-tree (Payer), 3.5%	Dominant (US$ 100,000)	None
14	Li et al. (2017)^42 (US)	Multigene testing of women with high-risk of breast or ovarian cancer (BRCA testing)	NGS of BRCA1, BRCA2, TP53, PTEN, CDH1, STK11 & PALB2; 14.9%; US$2,418	No	RRM (42%)	Markov (Payer), 3%	78,118/QALY (US$ 100,000)	Cost of testing, probability of variant positive, prophylaxis uptake
15	Eccleston et al. (2017)^43 (UK)	Women with unselected ovarian cancer (No testing)	NGS of BRCA 1 & 2; 13%; US$440	Yes, first & second-degree relatives	RRM (50%), RRSO (75%)	Microsimulation (Payer), 3.5%	7,167/QALY (US$ 28,736 or £20,000)	None
16	Manchanda et al. (2015)^44 (UK)	Population-wide of healthy Ashkenazi Jewish women ≥ 30 years (Testing in high-risk)	TVA of BRCA 1 & 2 founder mutations; 2.5%; US$71	No	RRM (52%), RRSO (55%)	Decision-tree (Payer), 3.5%	Dominant (US$ 100,000)	None
17	Green et al. (2014)^45 (US)	Women with previous benign breast biopsy (No testing)	TVA of seven SNP mutations; 1.2–1.7%; US$945	No	Tamoxifen (20%)	Microsimulation (Payer), 3%	59,708/QALY (US$ 100,000)	Cost of testing, prophylaxis uptake, utility weights of outcomes
18	Kwon et al. (2010)^46 (US)	Women with unselected ovarian cancer (No testing)	NGS of BRCA 1 & 2; 10%; US$3,000	Yes, first-degree relatives	RRM (30%), RRSO (70%)	Markov (Societal), 3%	39,835/QALY (US$ 50,000)	Cascade testing, prophylaxis uptake
19	Holland et al. (2009)^47 (US)	Healthy women ≥ 35 years with high-risk (No testing)	NGS of BRCA 1& 2; 10%; US$2,542	No	RRM (15%), RRSO (25%)	Semi-Markov (Societal), 3%	11,699/QALY (US$ 50,000)	Prophylaxis uptake, utility weights for outcomes
20	Balmana et al. (2004)^48 (Spain)	Women ≥ 30 years with high-risk (No testing)	TVA of BRCA 1 & 2; 20%; US$546	Yes, first-degree relatives	RRM (9%)	Decision-tree (Not provided), not provided	8,337/QALY (Not provided)	Probability of variant positive
21	Tengs and Berry (2000)^49 (US)	Population-wide of healthy women ≥ 30 years with varying risk of breast or ovarian cancer (No testing)	NGS of BRCA 1 & 2; 0 population-wide: 0.8%; slightly high-risk: 10%; moderately high-risk: 20%; very high-risk: 50%; US$2580	No	RRM (100%), RRSO (100%)	Markov (Societal), 3%	Pop-wide: 2.5 million/QALY; slightly high-risk: 52,875/QALY; moderately high-risk: 23,327/QALY; very high-risk: 5,443–7,620/QALY (US$ 50,000)	None
22	Grann et al. (1999)^50 (US)	Population wide of Ashkenazi Jewish women ≥ 30 years (No testing)	TVA of BRCA 1 & 2 founder mutations; 2.5%; US$450	No	RRM (100%), RRSO (100%)	Markov (Payer), 3%	33,749/LYG (US$ 50,000)	Cost of testing, prophylaxis uptake

Abbreviations. CPM, contralateral prophylactic mastectomy; GDP, gross-domestic product; ICER, incremental cost-effectiveness ratio; LYG, life years gained; MM, mammogram; NGS, next generation sequencing; PARP, poly-adenosine-diphosphate ribose polymerase; QALY, quality-adjusted-life years; RRM, risk-eduction mastectomy; RRS, risk-reduction surgery; RRSO, risk-reduction salpingo-oophorectomy; TVA, targeted variant analysis; WTP: willingness-to-pay.

*Converted to 2021 US $, using the CCEMG-EPPI-Centre Cost Converter³⁰.

Table 2. Characteristics of economic studies on germline testing for colorectal cancer.

#	Study with publication year (Setting)	Target Population for germline testing (Referent strategy)	Test strategy, genes assessed, variant prob%; cost of test	Cascade Testing	Prophylaxis (frequency of surveillance in test positive vs. referent population)	Model (Perspective), discounting %	ICER (WTP threshold) in US$*/QALY or LYG	Sensitivity analysis, i.e. factors affecting cost-effectiveness
1	Guzauskas et al. (2022)^17 (US)	Population- wide of healthy adults (Testing high-risk)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2, 0.3%; US$200	Yes, first-degree relatives	Annual colonoscopy starting at 20 years vs. Decennial colonoscopy staring at 45 years	Markov (Payer), 3%	132,200/QALY (US$ 150,000)	Cost of testing, WTP threshold
2	Salikhanov et al. (2021)^23 (Switzerland)	Newly-diagnosed colorectal cancer patients (Reflex testing)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2; 3.0%; US$2,852	Yes, first-degree relatives	In relatives: Biennial colonoscopy starting at 25 years vs. Decennial colonoscopy starting at 50 years	Markov (Payer), 3%	54,122/QALY (US$ 83,190 or CHF 100,000)	Cost of testing, cascade testing
3	Ramzdan et al. (2021)^22 (Malaysia)	Convenience sample of colorectal cancer (No testing)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2; not provided; US$976	No	In relatives: Biennial colonoscopy vs. FIT ± colonoscopy starting at 50 years	Decision-tree (Payer), 3%	Dominant (Not provided)	None
4	Pastorino et al. (2020)^21 (Italy)	Newly-diagnosed colorectal cancer patients (No testing)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2; 2.8%; US$348	Yes, first-degree relatives	In relatives: Biennial colonoscopy with Aspirin ± endometrial sampling ± RRSO starting at 25 years vs. colonoscopy every 5 years starting at 45 years	Markov (Payer), 3.5%	2,388/QALY (US$ 42,599 or €30,000)	Probability of variant positive, cascade testing
5	Kang et al. (2020)^18 (Australia)	Newly-diagnosed colorectal cancer patients (No testing)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2; 2.8%; $832	Yes, first-degree relatives	In relatives: Annual colonoscopy vs. No colonoscopy	Microsimulation (Payer), 5%	46,078/LYG (US$ 34,770 or AU$ 50,000)	Assumed benefits of prophylaxis
6	Pereira et al. (2019)^51 (Portugal)	Healthy adults with high-risk i.e. ≥ 5% (No testing)	SNPs in COX-2, PGE2; 30%; US$49	No	Increased colonoscopy, frequency not provided vs. no colonoscopy	Decision-tree (Societal), 3%	76,750/QALY (US$ 50,000)	Cost of testing, benefits of prophylaxis
7	Chen et al. (2016)^52 (Taiwan)	Newly-diagnosed colorectal cancer patients (No testing)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2; 2.3%; US$1,000	Yes, first-degree relatives	In relatives: Biennial colonoscopy starting at 20 years vs. FIT ± colonoscopy starting at 50 years	Markov (Payer), 3%	169,887/LYG (US$ 50,000)	None
8	Snowsill et al. (2015)^53 (UK)	Newly-diagnosed colorectal cancer patients (No testing)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2; 8.4%; US$1,168	Yes, first-degree relatives	In relatives: Biennial colonoscopy starting at 20 years ± endometrial sampling ± RRSO vs. no colonoscopy	Decision-tree (Payer), 3.5%	15,544/QALY (US$28,736 or £20,000)	Cost of testing
9	Severin et al. (2015)^54 (Germany)	Patients with colorectal cancer; stage non-specific (No testing)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2; 2.8%; US$5,985	Yes, first-degree relatives	In relatives: Annual colonoscopy with Aspirin vs. Decennial colonoscopy starting at 55 years	Decision-tree (Payer), 3%	981,494/QALY (US$75,616 or £20,000)	Prophylaxis uptake
10	Gallego et al. (2015)^55 (US)	Patients referred to a genetics clinic for colorectal cancer screening (Reflex testing)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2, EPCAM, APC, BMPR1A, SMAD4, STK11, MUTYH, PTEN, TP53, CDH1; 15.4%; US$2,700	Yes, first, second, & third-degree relatives	In positive individuals: Increased colonoscopy, details not provided vs. no colonoscopy in test negative	Decision-tree (Payer), 3%	41,216/QALY (US$100,000)	Penetrance of tested variants
11	Barzi et al. (2015)^56 (US)	Population-wide of healthy adults & newly-diagnosed colorectal cancer patients (No testing)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2; 0.2% in general population and 3.0% in colorectal cancer patients; $3,520	Yes, first-degree relatives	Annual colonoscopy starting at 20 years vs. No colonoscopy	Markov (Societal), 3%	1.1 million/QALY (US$50,000)	Cost of testing
12	Wang et al. (2012)^57 (Singapore)	First-degree relatives of hereditary colorectal cancer patients (No testing)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2; 0–2.3%; US$1,962	No	Biennial colonoscopy starting at 21 years vs. Decennial colonoscopy starting at 50 years	Markov (Payer), 3%	Dominant (US$72,069 or 1× GDP)	Prophylaxis uptake
13	Ladabaum et al. (2011)^58 (US)	Newly-diagnosed colorectal cancer patients (No testing)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2; 3%; US$4,288	Yes, first-degree relatives	In relatives: Annual colonoscopy starting at 21 years ± endometrial sampling at 35 years ± RRSO vs. no colonoscopy	Decision-tree & Markov (Third-party), 3%	61,150/QALY (US$100,000)	Cascade testing
14	Dinh et al. (2011)^59 (US)	Population-wide of healthy adults with varying risk; 0–10% (Testing high-risk)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2; 5%; US$4,307	Yes, first & second-degree relatives	Annual colonoscopy + endometrial sampling starting at 25 years vs. Decennial colonoscopy starting at 50 years	Microsimulation (Societal), 3%	54,884/QALY (US$50,000)	Probability of variant positive, prophylaxis uptake, discount rate, WTP threshold
15	Mvundura et al. (2010)^60 (US)	Newly-diagnosed colorectal cancer patients (No testing)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2; 3%; US$3,457	Yes, first-degree relatives	In relatives: Annual colonoscopy starting at 20 years vs. Decennial colonoscopy staring at 50 years	Decision-tree (Payer), 3%	180,119/LYG (US$50,000)	Cost of testing, probability of variant positive, prophylaxis uptake
16	Olsen et al. (2007)^61 (Denmark)	Families with moderate- to high-risk of HNPCC (No testing)	NGS of Lynch syndrome variants, MLH1 & MSH2; 22%; US$153	No	Biennial colonoscopy starting at 25 years vs. no colonoscopy	Markov (Payer), 5%	353/LYG (Not provided)	Cost of prophylaxis (colonoscopy)
17	Neilsen et al. 2007^62 (Netherlands)	Families of MAP patients (No testing)	NGS of MAP; 1.5%; US$754	No	In family members: Biennial colonoscopy starting at 25 years vs. FOBT ± colonoscopy starting at 50 years	Markov, (Societal), 4%	37,972/QALY (US$93,458 or €80,000)	Cost of testing, probability of variant positive, discount rate
18	Breheny et al. (2006)^63 (Australia)	Healthy first-degree relatives of known FAP & HNPCC mutation carriers (No testing)	NGS of Lynch syndrome variants, MLH1, MSH2, MSH6, PMS2 & FAP; 50%: US$1,580	No	Lynch positive: Annual colonoscopy starting at 25 years and colorectal surgery at 45 years FAP positive: Sigmoidoscopy every 18 months starting at 10 years and colorectal surgery at 21 years	Markov (Payer), 5%	Dominant (Not provided)	Prophylaxis uptake

Abbreviations. FAP, familial adenomatous polyposis; FIT, faecal immunochemical test; GDP, gross-domestic product; ICER, incremental cost-effectiveness ratio; IHC, immunohistochemistry test; iFOBT, immunochemical faecal occult blood test; LYG: life years gained; MAP, MUTYH-associated polyposis; NGS, next generation sequencing; QALY, quality-adjusted-life years; WTP: willingness-to-pay.

*Converted to 2021 US$, using the CCEMG-EPPI-Centre Cost Converter³⁰.

Germline testing in HBOC: Cost-effective across most settings (Table 3)

The bulk of breast or ovarian cancer studies were on targeted testing of healthy women with higher-than-average risk (40.9%, n = 9 studies) and patients with pre-established breast or ovarian cancer diagnosis (36.4%, n = 8 studies). The analyses were typically performed from a health payer perspective (68.2%, n = 15 studies), included cascade testing of relatives of probands (40.9%, n = 9 studies), tested exclusively for BRCA variants (72.7%, n = 16 studies), had testing costs ≤ US$500 (59.1%, n = 13 studies), used Markov modelling (45.4%, n = 10 studies), and appraised the cost-effectiveness of testing against willingness-to-pay thresholds of US$30,000–100,000/QALY or LYG (90.9%%, n = 20 studies).

Germline testing was cost-effective or cost-saving in the vast majority of studies (90.9%, n = 20 studies, ICERs in US$: dominant = 91,392/QALY; 8,337–59,708/LYG). The two exceptions where testing was not cost-effective included the study by Tengs and Berry⁴⁹, that assessed population-wide testing in the general population (BRCA prevalence: 0.8%, ICER: US$2.5million/QALY), and the study by Saito et al.²⁶ in women with metastatic breast cancer, where testing was performed to inform treatment with the PARP inhibitor, olaparib (ICER: US$158,630/QALY).

Evaluating the cost-effectiveness of germline testing for HBOC across specific settings, the ICERs of the testing strategy compared to the referent (no germline testing or testing limited to those with family history/clinical risk) for (1) population-wide testing of healthy women was US$344 − 2.5 million/QALY; (2) healthy women at high-risk was US$dominant = 78,118/QALY; 8,337/LYG; (3) healthy women with previous benign breast biopsy was US$59,708/LYG; (4) women with existing breast or ovarian cancer was US$3,012–72,566/QALY, 39,835/LYG; and (5) women with metastatic breast cancer was US$158,630/QALY.

Sensitivity analyses revealed that the uptake of risk-reduction surgeries (45.5%, n = 10 studies), cost of germline testing (36.4%, n = 8 studies), the proportion of test positive cases (18.2%, n = 4 studies), and the assumed survival benefits or utility of testing (13.6%, n = 3 studies) were some of the major factors influencing cost-effectiveness.

Germline testing for colorectal cancer: cost-effective in people with high-risk (Table 4)

Studies in the colorectal cancer group assessed the cost-effectiveness of germline testing in people with high-risk (38.9%, n = 7 studies) and those with established colorectal cancer (50.0%, n = 9 studies). Few studies evaluated the economics of population-wide testing (11.1%, n = 2 studies). Most studies were designed from a health payer perspective (72.2%, n = 13 studies), included a cascade testing component (66.7%, n = 12 studies), tested for lynch syndrome variants, i.e. MLH1, MSH2, MSH6, PMS2 (88.9%, n = 16 studies), had testing costs > US$1000 (55.6%, n = 10 studies), used Markov models (50.0%%, n = 9 studies), and had WTP thresholds between US$30,000 and 100,000/QALY or LYG (77.8%, n = 14 studies).

Overall, the results for cost-effectiveness in colorectal cancer studies were mixed. While germline testing was cost-effective in healthy people with high-risk (ICERs: US$32,322–76,750/QALY; dominant-353/LYG), there was widespread variation in findings for colorectal cancer patients (ICERs: US$dominant = 54,122/QALY; 98,790–6.3 million/LYG) and in population-wide testing scenarios (ICERs: US$132,200/QALY; 1.1 million/LYG).

Among studies that were not cost-effective, four^52,54,56,60 out of five studies^{18,52,54,56,60} had testing costs above US$1,000. In the 2015 study by Barzi et al.⁵⁶, the authors recommended against universal germline testing and expressed that the low prevalence of variants associated with colorectal cancer in the general population (0.2%) and substantial costs of testing (US$3,520) explain their observed findings. Meanwhile the more recent, 2022 study by Guzauskas et al.¹⁷ saw a substantial drop in testing costs (US$200) and the ICER for population-wide testing versus testing those with high-risk was US$132,200/QALY. The authors concluded that universal germline testing was likely to be cost-effective after using a WTP cut-off of US$150,000/QALY. Overall, cost-effectiveness of germline testing in colorectal cancer was sensitive to the price of testing (38.9%, n = 7 studies), compliance rates of surveillance (27.8%, n = 5 studies), probability of being variant positive (27.8%, n = 5 studies), the inclusion of cascade testing (15.8%, n = 3 studies), health benefits or utility of testing (11.1%, n = 2 studies), WTP thresholds (11.1%, n = 2 studies), and discounting (11.1%, n = 2 studies).

Summary of reporting quality (Supplementary Tables S3 and S4)

More than 77% of HBOC studies (n = 17 studies) provided information on at least 24 of the 28 CHEERS items. None of the studies had a health economic analysis plan (HEAP) with 58-core items as advised by CHEERS³¹. Engagement with patients and others was also another area that was grossly under-reported, with 95.4% of the studies failing to include information on the two items assessing this component. Other areas with under-reporting included title (4.6%, n = 1 study), abstract (4.6%, n = 1 study), study perspective (4.6%, n = 1 study), measurement of outcomes (18.2%, n = 4 studies), valuation of resources and costs (4.6%, n = 1 studies), currency and price conversion (4.6%, n = 1 study), characterization of heterogeneity (9.1%, n = 2 studies), study parameters (4.6%, n = 1 study), source of funding (10.0%, n = 2 studies), and details on conflict of interest (27.3%, n = 6 studies).

Similarly, 83.3% of colorectal cancer studies (n = 15 studies) reported on at least 24 of the 28 CHEERS items. Information on HEAP (100%, n = 19 studies), approach to engagement with patients and others (100%, n = 19 studies), effect of engagement with patients and others (100%, n = 19 studies), measurement of outcomes (38.9%, n = 7 studies), valuation of resources and costs (16.7%, n = 3 studies), characterizing heterogeneity (11.1%, n = 2 studies), source of funding (11.1%, n = 2 studies), and statement on conflict of interest (5.6%, n = 1 study) were under-reported.

Discussion

In the current systematic review, we examined previous economic evaluations on the cost-effectiveness of testing for germline variants associated with breast, ovarian, or colorectal cancer. Germline testing was cost-effective in the majority of hereditary breast or ovarian cancer (HBOC) studies (20 of 22 studies) and approximately three-quarters of colorectal cancer studies (13 of 18 studies). Specifically, in the HBOC group, germline testing was likely to be cost-effective in (1) population-wide testing of all healthy women (variant prevalence of 0.5–0.7%); (2) women with higher-than-average risk (variant prevalence ≥ 2.5%); and (3) women with existing breast or ovarian cancer (variant prevalence of 5.5–17.2%), preferably with cascade testing of relatives of probands. The situation is more complex in colorectal cancer where testing was cost-effective when performed in people with higher-than-average risk (lynch-syndrome variant prevalence ≥ 1.5%) but displayed mixed results with widespread variation when performed in patients with existing colorectal cancer. Incidentally we also explored the economics of germline testing in prostate cancer, another cancer group associated with a high prevalence of heritable mutations^64,65. The National Comprehensive Cancer Network (NCCN) recommends germline testing based on family history or clinical stage/risk group of prostate cancer (high risk localized, very-high localized, regional node positive, and metastatic prostate cancer)⁶⁶. It is anticipated that these broad recommendations would impose a significant workload on the health care system⁶⁷. The benefits of germline testing in such a wide audience is unclear^68,69, except for the niche segment of men with metastatic castration resistant prostate cancer (mCRPC), where testing was used to guide personalized treatment with PARP inhibitors and improved survival^70,71. The economics of adopting such an approach, however, is not clear. Our search strategy led us to two studies^19,24 on the cost-effectiveness of genetic test-directed treatment with the PARP inhibitor, Olaparib in men with mCRPC. However, the setting for both studies were men who tested positive for three or 15 pre-specified gene alterations, i.e. both the intervention group (olaparib therapy) and the comparator (standard care) were men who were variant positive. It was not possible to examine the cost-benefits of a germline testing versus no testing scenario and both studies^19,24 were excluded from the review. Overall, the evidence on cost-effectiveness of germline testing in prostate cancer is limited, therefore policy advice for testing in this cancer group is not yet possible.

Sensitivity analyses performed within the final sample of studies included in the review showed that the cost-effectiveness of germline testing could be influenced by the prevalence of tested variants, cost of testing, uptake of prophylactic measures (risk reduction surgeries, intensive surveillance), the survival benefits associated with prophylaxis, WTP thresholds, and cost inflation. Differences in real-world data or assumptions made about these parameters may have contributed to the observed heterogeneity in ICERs across studies.

Examining each of the influential parameters, 36–37% of studies in both HBOC and colorectal cancer reported their results as being sensitive to cost of germline testing. Pricing was on average higher in colorectal cancer relative to HBOC. Around 36.8% of colorectal cancer studies reported testing costs over US$2,000, a stark contrast to 18.2% of studies with similar pricing in HBOC. A few recent studies^{17,20,21,25,32,34} saw a reduction in price of testing, yet for the most part costs remained highly variable across settings^27,72.

Assumptions on uptake of prophylactic risk reduction surgeries also influenced results and was the most predominant factor affecting cost-effectiveness in HBOC studies. Uptake rates for RRM were usually lower than those for RRSO. The variation in adoption of RRM (9–100%) and RRSO (25–100%) rates across studies could be attributed to cultural differences in perceived ceilings for childbearing age, stigmas associated with the surgeries, costs of surgery, and patient perceptions on long-term survival post-surgery⁷³. In colorectal cancer where preventive surgeries have no clear benefits⁷⁴, elective risk reduction surgeries (e.g. colectomy) were not the mainstay of management and test positive cases were usually followed-up with intensive surveillance (colonoscopy or sigmoidoscopy every 1–3 years). The study by Breheny et al.⁶³ was the sole exception where intensive surveillance was followed by proactive colorectal surgery (at 45 years for Lynch syndrome carriers; at 21 years for familial adenomatous polyposis). Elective colorectal surgery could reduce the incidence of metachronous colorectal cancer in younger patients with high-risk^75,76; however, the strategy is not widely employed and may need further investigation⁷⁷.

Apart from the differences in costs of testing and assumptions on prophylaxis, studies also differed in terms of the modelling techniques used for their analysis. Cohort-based state-transitions models such as the Markov were the most prominent design used in both HBOC (45% of studies) and colorectal cancer studies (52.6% of studies). Decision-tress (HBOC: 22.7%; colorectal: 31.6%) and patient-level microsimulation models (HBOC: 27.3%, colorectal: 10.5%) were the other two designs used. Markov models may have been used more frequently owing to their less stringent computation and individual patient-level data requirements. Unlike the memory-less Markov models, microsimulation models have the advantage of being able to account for patient-level characteristics and their effect on future events (such as uptake of prophylaxis, etc.) and are, therefore, a better design for screening type interventions⁷⁸. However, they do require more extensive individual patient data and are also computationally more intensive. Finally, decision trees are logic-based designs lacking the ability to account for time-dependent variables or recurring events and are, therefore, the simpler of the three modelling techniques⁷⁹.

Germline testing is also useful to inform treatment with drugs that act on underlying pathogenic mechanisms for hereditary cancer. The cost-effectiveness of germline testing guided therapy was assessed in two studies^26,45. Saito et al.²⁶ examined the cost-effectiveness of the PARP inhibitor, olaparib, in BRCA positive women with metastatic breast cancer versus standard chemotherapy in those without mutation profiling. Testing-guided treatment was not cost-effective (ICER: US$158,630 QALY) and the authors commented that cost-effectiveness could improve with a reduction in costs of BRCA profiling and olaparib therapy. Meanwhile, Green et al.⁴⁵ examined the utility of germline testing to determine eligibility for tamoxifen therapy in women with previous benign breast biopsy (i.e. high-risk of future breast cancer) and demonstrated cost-effectiveness (ICER: 59,708/QALY).

Our findings for the cost-effectiveness of germline testing in HBOC and colorectal cancer concur with previous reviews^13–15 in the utility of targeted screening of people with higher-than-average risk. However, unlike the previous work by Koldekoff et al.¹⁴ where studies on population-wide genetic testing for HBOC were excluded, we made no such omission. Our decision to include these studies holds merit, as recent studies^17,20,32,34 suggest that population-wide testing could be cost-effective in HBOC as well as colorectal cancer. These findings could be explained by reduction in germline testing costs and the choice of WTP limits. All four of the recent studies^17,20,32,34 had testing costs ≤ US$200, a reduction from the typical range in the review (Tables 3 and 4). Regarding WTP thresholds the three HBOC studies^20,32,34 used the more commonly used cut-off of US$100,000/QALY or LYG, while the colorectal cancer study¹⁷ based its conclusion at moderately high WTP of $150,000 per QALY. Other notable observations from the current review were the heterogeneity among studies both in modelling techniques (Markov, microsimulation, decision-tree, etc.), uptake of prophylactic measures, compliance with intensive surveillance, probability of being variant positive, discounting, and WTP thresholds. Differences in most of these variables may, however, be an inherent tenet of the population or setting, induced by cultural perceptions towards the perceived value of testing or dependent on the GDP of countries, etc. Extending the discussion on variation in findings across country or setting, Manchanda et al.³² assessed the cost-effectiveness of BRCA testing across multiple countries. Population-wide BRCA testing was cost-saving in high-income countries (UK, US, and Netherlands, ICERs: dominant, WTP: US$42,656–57,589/QALY), cost-effective in upper-middle-income countries (China and Brazil, ICERs: US$13,579–18,066/QALY; WTP: US$15,181–15,531/QALY), and cost-prohibitive in lower-middle income countries (India, ICER: US$23,031/QALY, WTP: US$6,574/QALY). Overall, due to the heavy influence of individual-level variables, patient level-microsimulation may be a more appropriate design to evaluate cost-effectiveness of germline testing than the widely used cohort level Markov models.

Table 3. Summary of characteristics of economic studies on germline testing for breast or ovarian cancer.

Study characteristics	Breast or ovarian cancer (n = 22 studies)
Population/setting
Multiple countries	5 (22.7%)
US	7 (31.8%)
UK	2 (9.1%)
Asia	3 (13.6%)
Australia	1 (4.6%)
Others	4 (18.2%)
Initial target populations for germline testing
Population-wide	5 (22.7%)
Targeted testing
Healthy women at high-risk	9 (40.9%)
Breast or ovarian cancer patients	8 (36.4%)
Cascade testing
Included in the model	10 (45.4%)
Germline variants tested
BRCA1	1 (4.6%)
BRCA 1 & 2	15 (68.2%)
BRCA & others	6 (27.3%)
Cost of germline testing*
≤ US$200	2 (9.1%)
US$201–500	11 (50.0%)
US$501–1,000	4 (18.2%)
≥ US$1,000	5 (22.7%)
Study perspective
Health payer	15 (68.2%)
Societal	3 (13.6%)
Health payer & societal	3 (13.6%)
Not provided	1 (4.6%)
Modelling design
Decision-tree	5 (22.7%)
Markov	10 (45.4%)
Semi-Markov	1 (4.6%)
Microsimulation	6 (27.3%)
Time horizon
5 years or less	1 (4.6%)
lifetime	21 (95.4%)
Discount rate
≤ 2%	2 (9.1%)
3%	12 (54.6%)
3.5%	6 (27.3%)
4–5%	2 (9.1%)
Willingness-to-pay (WTP) threshold, US$
≤50,000	8 (36.4%)
51,000–100,000	12 (54.6%)
101,000–150,000	1 (4.6%)
Not provided	1 (4.6%)
ICERs, stratified by target population for testing vs. referent strategy in US$*
Population-wide	US$344 − 2.5 million/QALY
Targeted testing
Healthy women with high-risk	Dominant: US$78,118/QALY; 8,337/LYG
Women with previous benign breast biopsy	59,708/LYG
Women with breast or ovarian cancer	US$3,012–72,566/QALY; 39,835/LYG
Metastatic breast cancer	US$158,630/QALY
Results of economic evaluations
Cost-saving	3 (13.6%)
Cost-effective	17 (77.3%)
Not Cost-effective	2 (9.1%)
Factors influencing results (Sensitivity analysis)
Cost of testing	8 (36.4%)
Uptake of prophylaxis	10 (45.5%)
Probability of being variant positive	4 (18.2%)
Assumed benefits of prophylaxis	3 (13.6%)
Cascade testing	2 (9.1%)
Discounting	2 (9.1%)

*Converted to US$ using the CCEMG-EPPI-Centre Cost Converter³⁰.

Table 4. Summary of characteristics of economic studies on germline testing for colorectal cancer.

Study characteristics	Colorectal cancer (n = 18 studies)
Population/setting
US	6 (33.3%)
UK	1 (5.6%)
Asia	3 (16.7%)
Australia	2 (11.1%)
Others	6 (31.6%)
Initial target populations for germline testing
Population-wide	2 (11.1%)
Targeted testing
People with high-risk	7 (38.9%)
Colorectal cancer patients	9 (50.0%)
Cascade testing
Included in the model	12 (66.7%)
Germline variants tested
Lynch	14 (77.8%)
Lynch & others	2 (11.1%)
Other colorectal cancer variants	2 (11.1%)
Cost of germline testing*
≤ US$200 or less	2 (11.1%)
US$201–500	2 (11.1%)
US$501–1,000	4 (21.2%)
US$1,000–2,000	3 (16.7%)
≥US$2,000	7 (38.9%)
Study perspective
Health payer	13 (72.2%)
Societal	4 (22.2%)
Third party	1 (5.6%)
Modelling design
Decision-tree	6 (33.3%)
Markov	9 (50.0%)
Microsimulation	2 (11.1%)
Cross-sectional	1 (5.6%)
Time horizon
lifetime	16 (88.9%)
not provided	2 (11.1%)
Discount rate
3%	12 (66.7%)
3.5%	2 (11.1%)
4–5%	4 (22.2%)
Willingness-to-pay (WTP) threshold in US$/QALY or LYG
≤ 50,000	8 (44.4%)
51,000–100,000	6 (33.3%)
101,000–150,000	1 (5.6%)
not provided	3 (16.7%)
ICERs, by target population for testing vs. referent population in US$*
Population-wide testing	132,000/QALY; 1.1 million/LYG
Targeted testing
Healthy adults with high-risk	32,322 − 76,750/QALY; dominant: 353/LYG
Colorectal cancer patients	dominant: 54,122/QALY; 98,790–6.3 million/LYG
Results of economic evaluations
Cost-saving	3 (16.7%)
Cost-effective	10 (55.6%)
Not cost-effective	5 (27.8%)
Factors influencing results (Sensitivity analysis)
Cost of testing	7 (38.9%)
Compliance with increased surveillance	5 (27.8%)
Probability of being variant positive	5 (27.8%)
Assumed benefits of increased surveillance	2 (11.1%)
Cascade testing	3 (15.8%)
WTP threshold	2 (16.7%)
Discounting	2 (11.1%)

*Converted to US$ using the CCEMG-EPPI-Centre Cost Converter³⁰.

Our assessment of reporting quality using CHEERS³¹ suggested that studies need to do a better job at reporting measurement of outcomes. Patient or stakeholder engagement was another glaring omission in most studies. Items requiring reporting in these areas have been introduced in the recently updated CHEERS criteria³¹. This is an important area of research, especially in interventions for cancer patients, where there are uncertainties about WTP limits^80–83. Future studies could provide valuable insight by weighing patient and stakeholder preferences for cost-utility of germline testing compared to other healthcare needs.

The limitations of this review are noteworthy. First, we assessed the cost-utility of germline testing and did not include evaluations on genomic or somatic testing. Both testing modalities fall under the purview of genetic testing, yet there are clear distinctions between them. While somatic testing looks for genetic changes that develop over the individual’s lifetime, germline testing is focused on heritable mutations and our findings are limited to the latter. Second, the heterogeneity of included studies regarding the target group for testing and the small number of articles within each of these strata prevented us from performing a metanalysis. Also, our findings are not comprehensive as we did not include studies published in non-English languages. Third, information on prognostic markers such as HER2 or triple negative status (additionally provides information about estrogen or progesterone receptors on cancer cells) was not provided in most HBOC studies. Patients with BRCA mutations have a higher likelihood of being triple negative⁸⁴, which may further reduce their long-term survival. Information on triple negative status could, therefore, explain some of the variation in cost-effectiveness across studies.

Germline testing in breast, ovarian, or colorectal cancer has the potential to provide timely and cost-effective personalized healthcare^85,86. Previous reviews of HBOC^14,87 recommend testing based on family history and for women with pre-existing breast or ovarian cancer. Our evaluation considers recently published studies on population-wide testing^20,32,34 and suggests that testing for HBOC is likely to be cost-effective across most scenarios, i.e. population-wide testing of healthy women, targeted testing of those with family history, and testing of women with existing breast or ovarian cancer. In colorectal cancer where genetic sequencing costs were relatively high and pathogenic variant prevalence lower, germline testing demonstrated cost-effectiveness in people with higher-than-average risk but was inconclusive in other settings. Testing is also recommended to inform personalized treatment with PARP inhibitors in advanced stage breast, ovarian, and prostate cancer^1,3,66. Several clinical trials have demonstrated the significant survival benefits of PARP over the standard care option^88,89. Yet, studies on the economics of these co-dependent technologies (germline testing followed by PARP therapy)⁹⁰ have been limited and inconclusive. The approach is currently not financially feasible in HBOC^26,91, but is potentially cost-effective in prostate cancer after adopting relatively high WTP cut-offs of US$150,000/QALY²⁴. Despite the gain in QALYs the high price of PARP inhibitors may be the primary driver of the increased costs^24,26,89,91. Policy advice on germline testing should recommend weighing the benefits of testing with careful consideration to the prevalence of pathogenic variants in the testing scenario, costs of genetic sequencing and subsequent personalized treatments, and the uptake and benefits of prophylactic measures.

Transparency

Author contributions

ST performed the literature search, data abstraction and synthesis, quality assessment, and prepared the first draft of the manuscript. HT, PS, and SH contributed to the conceptual design of the study and towards revisions to the manuscript. BH reviewed the literature search, methodology, quality assessment, and provided feedback for revisions. All the authors read and agreed to the final version of the article.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgements

None reported.

Declaration of funding

This research is part of a doctoral thesis by Srinivas Teppala. Dr. Teppala is the recipient of a Griffith University Health Group International Postgraduate Scholarship and a Griffith University Postgraduate Research Scholarship (2019–2022). Haitham Tuffaha holds a Priority Impact Research Award – Future Leader funded by Prostate Cancer Foundation of Australia.

Declaration of financial/other relationships

The authors report no conflicts of interest related to this manuscript.