If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Corresponding author at: The Ohio State University Comprehensive Cancer Center, M210 Starling Loving Hall, 320 West 10th Avenue, Columbus, OH 43210, United States of America.
Upfront multi-gene panel testing (MGPT) for endometrial cancer (EC) improves the diagnosis of Lynch syndrome (LS).
•
MGPT for all newly diagnosed EC demonstrated $17.1 M cost savings per year relative to the current testing algorithm.
•
Incorporation of universal MGPT represents an opportunity to improve access to genetic counseling and testing.
Abstract
Objectives
The routine use of upfront universal germline genetic testing among patients with newly diagnosed endometrial cancer (EC) has been proposed to improve diagnosis of Lynch syndrome (LS) and discover pathogenic variants (PVs) in other cancer susceptibility genes. We propose an algorithm prioritizing upfront multi-gene panel testing (MGPT) for newly diagnosed EC patients.
Methods
A decision analysis compared the cost of the current algorithm of universal mismatch repair (MMR) immunohistochemistry (IHC) for all EC cases to a new MGPT algorithm that employs upfront MGPT for all EC cases and reserves MMR IHC for the recurrent setting. The increase in the number of LS diagnoses using upfront MGPT, and the number of patients with PVs in BRCA1 and BRCA2 are also estimated.
Results
The MGPT algorithm demonstrated a cost savings of $259 per patient. Assuming 66,950 new cases of EC per year, this would represent $17.1 M of cost savings per year. When applied to all new diagnoses of EC in one year, the MGPT algorithm identified 660 (1%) additional cases of LS that would have been missed with the current algorithm. An additional 660 (1%) EC patients with BRCA1 or BRCA2 PVs would be diagnosed only through implementation of universal MGPT.
Conclusions
The use of universal upfront MGPT is a practical consideration for patients with newly diagnosed EC for cost savings and improved diagnosis of highly penetrant cancer syndromes. Incorporation of germline genetic testing in the upfront setting represents an opportunity to improve access to genetic counseling and testing, and ultimately an avenue to achieve equity and improve the lives of our patients with EC and their families.
1. Background
Endometrial cancer (EC) is the most common gynecologic malignancy in the US [
]. Identification of LS in patients with EC affords cancer prevention opportunities for patients and their family members. The current process for making the diagnosis of LS often involves tumor screening with immunohistochemistry (IHC) for mismatch repair (MMR) proteins and reflex MLH1 promoter hypermethylation testing when indicated, followed by confirmatory germline genetic testing in patients with unexplained MMR deficiency in their tumor.
The use of tumor-based screening for LS in EC was extrapolated from its use in colorectal cancer (CRC) however important differences exist. For example, a greater proportion of EC will have MMRd (loss of MLH1 and PMS2) which is frequently explained by MLH1 promoter hypermethylation, underscoring the importance of reflex MLH1 promoter hypermethylation testing in the algorithm of tumor screening. CRC associated with LS is more common in patients under 50, but screening patients only under age 50 will miss the most common causes of LS in EC [
Combined microsatellite instability, MLH1 methylation analysis, and immunohistochemistry for lynch syndrome screening in endometrial cancers from GOG210: an NRG oncology and gynecologic oncology group study.
]. The high prevalence of MSH6 and PMS2 pathogenic variants (PV) causing LS in EC patients underscores the importance of a universal approach to screening over a family history-based referral as families harboring MSH6 or PMS2 PVs often do not represent the classic Amsterdam-criteria family [
Combined microsatellite instability, MLH1 methylation analysis, and immunohistochemistry for lynch syndrome screening in endometrial cancers from GOG210: an NRG oncology and gynecologic oncology group study.
]. A large, multi-institution study that prospectively tested newly diagnosed EC patients with tumor and germline testing found that the most common causes of LS in EC were PMS2 and MSH6 PVs and identified numerous challenges with IHC screening that ultimately could have resulted in missed diagnoses of LS [
]. It is important to note that even in those instances in which tumor screening identifies possible LS, the diagnosis is not confirmed until germline genetic testing is completed, with several studies demonstrating a significant drop off between an abnormal IHC result and completion of genetic testing [
Universal endometrial cancer tumor typing: how much has immunohistochemistry, microsatellite instability, and MLH1 methylation improved the diagnosis of lynch syndrome across the population?.
]. MGPT allows for improved diagnosis of LS compared to the tumor testing approach as well as the identification of other hereditary cancer syndromes. Levine et al. showed that 1% of an unselected EC cohort harbored BRCA PVs [
]. These findings together support the use of upfront MGPT for all EC patients. Here we propose an algorithm for testing in newly diagnosed EC patients, prioritizing upfront germline MGPT and reserving the use of MMR IHC in the recurrent setting. Cost analysis was performed to model the impact of implementation of this algorithm if applied to routine clinical practice in the US.
2. Methods
A decision analysis model was built using TreeAge Pro (Fig. 1) to compare two scenarios:
1)
Current algorithm: IHC of MMR proteins MLH1, MSH2, MSH6 and PMS2 is performed for all newly diagnosed EC cases. Tumors demonstrating loss of MLH1 and PMS2 undergo reflex MLH1 promoter hypermethylation testing. Those with loss of MLH1 and PMS2 without MLH1 promoter hypermethylation, or tumors with absent MSH2 and MSH6, MSH6 only or PMS2 only are referred for genetic testing.
2)
MGPT algorithm: germline MGPT is performed for all newly diagnosed ECs in place of universal IHC testing. MMR IHC is performed for patients who develop recurrence.
Fig. 1Decision analysis model used for evaluating cost savings of IHC-based screening (current algorithm) versus universal multi-gene panel testing (MGPT algorithm) for newly diagnosed endometrial cancer.
]. The rate of missed LS diagnoses with IHC systems in place was also estimated. In the MGPT algorithm, only recurrent cases (estimated to be 17% of cases [
]) would undergo MMR IHC to determine if immunotherapy is warranted and this would not include MLH1 promoter hypermethylation testing since MMR gene mutation status was already assessed by MGPT at the time of initial diagnosis. MMRd is expected in 29% of the EC population, with 3% of that attributable to LS [
]. The cost of IHC and MLH1 promoter hypermethylation was determined from costs at The Ohio State University Medical Center and is consistent with rates published in other cost-effective analyses of universal tumor testing [
The model provides an estimated cost per person for each testing strategy. This amount was multiplied by the number of anticipated new EC cases in 2022 [
], showing expected costs in a year. The expected number of LS cases and BRCA-associated cases that would be captured by the new algorithm was extrapolated based on the rates of LS, including cases missed by IHC, as well as the number of BRCA-associated cases that would not be identified by IHC [
The MGPT algorithm demonstrated a cost savings of $259 per person per year. With 65,950 estimated new cases in 2022, this represents a $17.1 M cost savings in a single year (Table 2). The expected rate of LS diagnoses missed by the current algorithm is 1% based on the OPTEC data. Assuming a 1% rate of missing LS by the current algorithm, the MGPT algorithm will identify 660 cases of LS that would have been missed with IHC alone.
Table 2Comparative costs of testing strategies.
Mean cost per person
Total cost
Assuming 65,950 new endometrial cancer diagnoses in one year (Siegel, 2022), all costs in US dollars.
Cost savings
Assuming 65,950 new endometrial cancer diagnoses in one year (Siegel, 2022), all costs in US dollars.
Current algorithm
$579
$38,185,050
–
MGPT algorithm
$320
$21,104,000
$1,70,81,050
Assuming 65,950 new endometrial cancer diagnoses in one year (Siegel, 2022), all costs in US dollars.
Similarly, the MGPT algorithm will identify an additional 660 patients carrying highly penetrant PVs in BRCA1 or BRCA2 based on performance of this panel [
]. A one-way sensitivity analysis demonstrated that if the cost of MGPT was increased from the baseline $250 to $513 per patient, the cost of the MGPT algorithm would be equivalent to the current algorithm, with the added benefit of identifying more patients with LS and BRCA1 and BRCA2 PVs (Fig. 2). If the cost of MGPT decreases to $100, cost savings per person increases from $259 per person to $407 per person, with a total cost savings to $27.2 M per year.
Fig. 2One-way sensitivity analysis varying the cost of multi-gene panel testing (MGPT) showing the effect of varying cost of MGPT on mean cost of testing strategy per patient for the current and MGPT algorithms. If cost of MGPT increases to $513 per test, the mean cost per patient would be equivalent in each testing strategy.
A tornado diagram showing the results of one-way sensitivity analyses of the cost and probability variables in our model show that the model is most sensitive to changes in the cost of IHC, the proportion of patients with dMMR on IHC, and the cost of MGPT (Fig. 3). Increases in IHC cost, higher levels of dMMR on IHC, and lower cost of MGPT increase the incremental cost savings of universal MGPT.
Fig. 3Tornado diagram showing the effect of varying model inputs on the incremental cost of IHC-based screening (current algorithm) versus universal multi-gene panel testing (MGPT algorithm). Each tier of the tornado diagram represents a one-way sensitivity analysis varying the input across a range of values. The model was most influenced by changes in the cost of IHC, the proportion of patients with mismatch repair deficient (dMMR) tumors on IHC, and the cost of MGPT. Increases in IHC cost, higher levels of dMMR, and lower cost of MGPT increase the incremental cost savings of universal MGPT. Cost variables presented in US dollars.
Here, we demonstrate that if the MGPT algorithm is expanded across all new EC cases in the US today, there is potential for cost savings of more than $17 M annually, with the added benefit of capturing LS diagnoses missed by IHC and finding PVs in other highly penetrant cancer susceptibility genes like BRCA1 and BRCA2.
The importance of identifying cancer patients with hereditary cancer predisposition syndromes has already been established and has led to the recommendations for universal genetic testing for patients with ovarian, fallopian tube, and primary peritoneal cancer [
]. This was also the impetus for implementing universal IHC with reflex MLH1 hypermethylation testing for all endometrial cancers. However, at the time of implementation of tumor screening, germline genetic testing examined single genes and was vastly more expensive. Delayed development of clinical testing for PMS2 presented additional challenges with regard to access to testing. With the advent of large multi-gene panel tests and significant decreases in the cost of testing, upfront universal germline testing presents an opportunity for more comprehensive genetic evaluation at a favorable cost profile.
The use of universal MGPT in the upfront setting for EC has been shown to find 10% likely pathogenic or pathogenic variants in cancer susceptibility genes in multiple EC cohorts [
]. Importantly, Levine et al. demonstrated that MGPT captured more LS diagnoses, with one-third of LS diagnoses identified as cases that would have been missed by existing systems [
]. Reasons for missed cases were false-negative IHC results, insufficient remaining sample for tumor testing, or lack of a universal IHC screening system with family histories that would not necessarily trigger a genetics referral [
Multiple studies show the challenges with a workflow that starts with tumor-based testing to make a diagnosis of LS. Kahn et al. performed a meta-analysis including 6649 patients with EC from ten different countries. Among 1011 patients with tumor-based indications for germline testing, only 893 (88%) completed germline genetic testing [
Universal endometrial cancer tumor typing: how much has immunohistochemistry, microsatellite instability, and MLH1 methylation improved the diagnosis of lynch syndrome across the population?.
]. In a study from the Netherlands, Tjalsma et al. showed that of 15 patients with tumor testing suggestive of LS, three (20%) were not informed of the tumor testing results and were not referred for germline genetic testing. Of the remaining 12 patients referred for genetic counseling, only nine completed testing leaving only 9/15 (60%) patients with likely LS having completed genetic testing [
]. Lee et al. describe that at their institution (NY, US), 10/583 (1.7%) patients with EC had germline genetic testing confirming the diagnosis of LS. A country-wide initiative in Canada found that despite universal tumor testing, only 0.8% of EC patients were confirmed to have LS [
]. With the expected rate of LS in an EC population of 3%, these studies show that even with universal tumor testing, LS is underdiagnosed. Further, a model by Gudgeon et al. shows that even with imperfect use of routine upfront germline genetic testing in EC patients, proceeding straight to germline genetic testing diagnoses more LS and was not more costly than a tumor-based approach. In that model, the rate of proceeding to genetic testing, or lack thereof in the tumor-based approach, made upfront germline genetic testing more effective for diagnosing LS [
Model-based re-examination of the effectiveness of tumor/immunohistochemistry and direct-to-sequencing protocols for lynch syndrome case finding in endometrial Cancer.
The expected rates of MMRd and germline PVs included in this model were derived from the prospective OPTEC multi-institutional trial in which the majority of EC patients were enrolled from non-NCI designated comprehensive cancer centers [
]. The fact that OPTEC is a state-wide initiative improves the generalizability of the findings, with the rates of PVs found in that study likely representative of what many other practice settings can expect. We used a genetic testing cost of $250, consistent with a recently published cost-effectiveness analysis looking at universal implementation of MGPT [
]. We recognize that despite innovations leading to a dramatic decrease in the cost of MGPT, a wide variation in cost still exists. Based on existing data, we believe $250 to be a reasonable estimate for this model. However, to account for the known variation in cost among genetic testing labs, we performed a one-way sensitivity analysis varying the cost of MGPT. We determined that if the cost of germline MGPT increased from $250 to $513 per test, the cost of universal upfront MGPT would be equivalent to the current strategy of IHC screening, with the added benefit of enhanced identification of patients with LS and BRCA PVs. This model also does not include the costs of genetic counseling given that the model only shows cost savings based on LS diagnosis and not all positive MGPT findings. By the current algorithm the expected percentage of patients referred to genetic counseling would be 6% based on expected mismatch repair abnormalities not explained by MLH1 promoter hypermethylation [
] plus patients with positive family history, multiple primary tumors, or other history based indications. Based on the expected 10% of patients with any positive MGPT result in an unselected EC population [
], ultimately the genetic counseling costs in each algorithm are similar however could be slightly higher in the MGPT algorithm.
This model does not account for the cost savings in preventing second cancers, nor does it account for the potential of successful cascade testing and prevention of cancer in at-risk relatives. Given that a large proportion of EC patients are cured, preventing second cancers is an important target for this population. Several studies have demonstrated low levels of colon cancer screening utilization among cancer survivors [
], placing greater risk for the development of second cancer when the diagnosis of LS is missed. Patients with LS have up to a 50% chance of developing a second primary cancer [
] with colorectal cancer being the most common. Colonoscopy screening has been shown to reduce incidence and mortality of colorectal cancer by 62% and 65% respectively [
]. Therefore, accurate diagnosis of LS is critical for improving outcomes for patients cured of EC. Similarly for BRCA carriers with up to an 85% lifetime risk for developing breast cancer, implementation of increased surveillance with breast MRIs and mammograms reduces mortality from breast cancer. Screening in LS and for BRCA carriers has previously been demonstrated to be cost-effective [
Beyond the costs savings demonstrated in this model, there are other important considerations when considering implementation of universal MGPT for EC patients. There are varying practice settings where patients with EC receive care. At our institution, the routine use of MMR screening in the upfront setting serves an additional role in determining eligibility for clinical trials. In places where MMR screening is used purely for the identification of LS, replacing this process with MGPT and reserving MMR testing for the recurrent setting may be a more practical approach. The implementation of upfront MGPT may be of particular importance in those centers that have not been able to establish universal MMR screening. Given the prognostic significance of the ProMisE algorithm [
] and the number of trials underway incorporating tumor molecular information into treatment decision making, it is likely that more comprehensive tumor testing will eventually become standard of care for EC patients. NCCN guidelines already encourage evaluation for POLE mutations, MMR, and p53 status to complement pathologic diagnosis [
], however current clinical applications in the upfront setting are limited. For some centers, expanded molecular testing may ultimately involve somatic tumor testing with outside genetic testing labs. Currently, genetic counselors are an integral part of coordinating this testing for ovarian cancer patients. Shifting to upfront germline MGPT for all ECs may facilitate the transition to the more routine use of tumor testing of ECs when the time comes.
The authors acknowledge the existing evidence demonstrating challenges in real-world implementation of universal germline genetic testing. It has now been greater than 10 years since guidelines recommended universal germline genetic testing for patients with epithelial ovarian cancer (EOC), and still, recent reports suggest that less than 50% of EOC patients complete germline genetic testing. Most studies point to practice setting, insurance status, and disparities in race as contributing factors [
]. Including the recommendation for upfront MGPT for all EC patients in national guidelines is an important step to overcoming barriers of insurance coverage and uncertainty about the utility of testing. Surmounting these barriers can facilitate the equitable delivery of care for EC patients across varying practice settings.
5. Conclusions
The use of upfront MGPT is a practical consideration for patients with newly diagnosed EC for cost savings and improved diagnosis of highly penetrant cancer syndromes. With the increasing integration of genetic information in gynecologic cancer care, the incorporation of universal germline genetic testing in the upfront setting represents an opportunity to improve access to genetic counseling and testing and ultimately, an avenue to improve the lives of our patients with EC and their families.
Author contributions
MDL and DAB conceived and designed the model, collected data, performed the analysis and drafted the original manuscript. DEC and PJG verified data and provided project supervision and administration. All authors participated in editing and reviewing the manuscript.
Funding
This research was funded by an Ohio State University James Comprehensive Cancer Center Statewide Pelotonia Cancer Impact Award and R01CA223219.
Declaration of Competing Interest
MDL, DEC, DAB and PJG have no disclosures.
HH is on the scientific advisory boards for Genome Medical, Invitae Genetics, and Promega. She is a consultant for 23andMe and GI OnDemand. She has stock/stock options in Genome Medical and GI OnDemand.
Combined microsatellite instability, MLH1 methylation analysis, and immunohistochemistry for lynch syndrome screening in endometrial cancers from GOG210: an NRG oncology and gynecologic oncology group study.
Universal endometrial cancer tumor typing: how much has immunohistochemistry, microsatellite instability, and MLH1 methylation improved the diagnosis of lynch syndrome across the population?.
Model-based re-examination of the effectiveness of tumor/immunohistochemistry and direct-to-sequencing protocols for lynch syndrome case finding in endometrial Cancer.
Assuming 65,950 new endometrial cancer diagnoses in one year (Siegel, 2022), all costs in US dollars.
00488-7&id=gr1.jpg){kind=link}
00488-7&id=gr2.jpg){kind=link}
00488-7&id=gr3.jpg){kind=link}