Cost-Effectiveness of Annual Prostate MRI and Potential MRI-Guided Biopsy After Prostate-Specific Antigen Test Results

Key Points Question Are an annual prostate magnetic resonance imaging (MRI) and potential MRI-guided biopsy cost-effective compared with a transrectal ultrasonographic (standard) biopsy for men with a positive prostate-specific antigen (PSA) test result? Findings In this economic evaluation, annual MRI and potential MRI-guided biopsy for prostate cancer found improved economic value compared with standard biopsy for men with a serum PSA level of 2.5 ng/mL or more beginning at 65 years of age. Meaning This study suggests that an annual MRI and potential MRI-guided biopsy are a cost-effective option from the federal payer perspective compared with standard biopsy for men beginning at 65 years of age for a serum PSA level of 2.5 ng/mL or more.


Introduction
Prostate cancer (PCa) is the most common noncutaneous cancer among men in the US and the second-highest cause of cancer deaths among men. 1,2There is evidence that significant cost savings can be achieved by enhanced precision in diagnosis. 3,4Magnetic resonance imaging (MRI) and MRI-guided biopsy have been introduced for the diagnosis of PCa.The standard for detecting PCa is transrectal ultrasonography (TRUS)-guided biopsy with 12 to 16 cores distributed throughout the prostate (standard biopsy). 5This approach is associated with risks of postbiopsy complications, 6 has limited diagnostic accuracy, 7 and is performed without visual confirmation of the cancer's location. 8e primary diagnostic limitations of standard biopsies are overdetection of clinically insignificant lower-grade PCa and underdetection or undergrading of clinically significant PCa. 9 There remains substantial controversy regarding prostate-specific antigen (PSA) screening for PCa, particularly its potential to result in overdiagnosis of low-risk disease.10 Despite these concerns, PSA screening remains a commonly used tool, and both American Urological Association and American Cancer Society guidelines continue to include PSA screening for PCa diagnosis.[11][12][13] Multiparametric MRI can assess which men have regions suspicious for cancer in the prostate, improving detection of clinically significant cancer and potentially avoiding biopsy when no target is evident.14 Despite proven efficacy, MRI and potential MRI-guided biopsy remain costly, and there is limited research evaluating the cost-effectiveness of this approach.5,15 We evaluate the cost-effectiveness of annual MRI and potential MRI-guided biopsy as a screening strategy for PCa detection compared with standard biopsy. W focused on newly eligible male Medicare beneficiaries who are 65 years of age over 10 years to maintain a model consistent with the initial phase of Medicare eligibility to the age limit beyond which PCa screening is generally not recommended.16

Model Overview
In this economic evaluation, a decision analytic Markov model based on current US epidemiologic and clinical data was constructed to evaluate the cost and effectiveness of 2 biopsy strategies: standard biopsy and MRI and potential MRI-guided biopsy in the US (eFigure 1 in Supplement 1).The model was constructed from a federal payer (Medicare) perspective and developed to evaluate outcomes based on prostate MRI-naive men entering the model at an age of 65 years over a 10-year time horizon, which is the earliest age that men undergoing PSA screening would be eligible for Medicare; nearly 66% of PCa cases are diagnosed among men aged 65 years or older, 17,18 while men older than 75 years are less likely to undergo definitive therapy for PCa because they face greater morbidity and mortality risk associated with other health issues. 19The study was conducted using simulated data and does not constitute human participants research and is therefore exempt from institutional review board approval.We followed the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) reporting guideline.
The Markov model uses annual transition probabilities to model a hypothetical individual's simulated history as they progress between multiple health states, factoring in health utilities and costs associated with treatment complications, mortality, and other relevant events (eFigure 1 in Supplement 1).At the end of each cycle (annual), costs and quality-of-life weights incurred in each health state are summarized and aggregated for each strategy.
The population was categorized by 4 initial PSA strata: less than 2.5 ng/mL, 2.5 to 4.0 ng/mL, 4.1 to 10.0 ng/mL, and more than 10.0 ng/mL (to convert to micrograms per liter, multiply by 1.0).Without loss of generality, the models assume that an individual's PSA category would not change over the 10-year horizon and have identical structures but with different probabilities of PCa recurrence due to disease progression.The aggressiveness of PCa is described by pathologic grade, stratified to grade groups based on the preponderance of histologic Gleason patterns. 20Patients with comorbidities such as diabetes, chronic obstructive pulmonary disease, congestive heart failure, or multiple comorbid conditions that can be exacerbated by treatment may defer PCa treatment out of concern that it would not improve length or quality of life. 23Those with high-grade PCa, however, may require treatment regardless of the presence of comorbidities. 24scriptive and structural validity was conducted between the economic and clinical authors first and then by the modeling team prior to evaluating final economic outcomes.Using the complete model, internal validity and consistency were checked via multiple sensitivity analyses by assessing the association of varying key parameter values, such as diagnostic test characteristics and cancer progression probabilities, with variation in relative Markov cohort sizes from one cycle to another.

JAMA Network Open | Oncology
Additional model details required to replicate this study are included in the eAppendix in Supplement 1.

Statistical Analysis Primary Measures
The incremental cost-effectiveness ratio (ICER) was used to assess cost-effectiveness.ICERs were calculated as the difference in costs divided by the difference in quality-adjusted life-years (QALYs) between the 2 competing strategies.All future costs and QALYs were discounted at a rate of 3% annually. 38We identified the efficient strategy for each PSA group using a standard willingnessto-pay (WTP) threshold of $100 000 per QALY; if the ICER was under $100 000 per QALY and the incremental QALY was positive, the strategy was considered cost-effective. 39

Sensitivity Analyses
A probabilistic sensitivity analysis was performed using Monte Carlo microsimulation with 10 000 trials to account for simultaneous uncertainty in model parameter values when calculating ICERs, and 1-way sensitivity analyses of key test characteristics (sensitivity and specificity of MRI and MRI-guided biopsy) were evaluated within these simulations.In addition, we generated tornado diagrams to describe the association of uncertainty of input parameters with incremental effectiveness and cost outcomes between strategies.All analyses were performed using TreeAge Pro, version 2021 R.12 (TreeAge Software LLC).Additional detail on primary measures is included in the eAppendix in Supplement 1.

JAMA Network Open | Oncology
Annual Prostate MRI and MRI-Guided Biopsy for Men With Positive PSA Results

Base-Case Scenario
The costs, QALYs, incremental values, and ICERs comparing the 2 strategies-the MRI and MRI-guided biopsy vs the standard biopsy-for the 4 different PSA strata are summarized in Table 3.
In the base-case scenario, the MRI and potential MRI-guided biopsy strategy was a cost-effective strategy compared with the standard biopsy for 3 of the 4 categories of PSA strata.The ICER was well below the WTP threshold of $100 000 per QALY for these 3 PSA strata, at $21 131 per QALY for a PSA of 2.5 to 4.0 ng/mL, $12 336 per QALY for a PSA of 4.1-10.0ng/mL, and $6000 per QALY for a PSA of more than 10.0 ng/mL.The ICER for a PSA less than 2.5 ng/mL was $187 558 per QALY.We found that the MRI and potential MRI-guided biopsy strategy improved in economic and clinical value as PSA increased.Although the total mean costs increased for both strategies for higher PSA strata, the incremental costs for the MRI and potential MRI-guided biopsy strategy decreased.and the probability for the PSA 4.1 to 10.0 ng/mL stratum was found using a mean probability of the PSA 4.1 to 6.0 ng/mL and PSA 6.1 to 10.0 ng/mL strata, and the values were calculated to annual probabilities.The probabilities of having adverse effects are the means of having adverse effects including pain, bleeding, hematuria, hemospermia, and rectal hemorrhage.The probabilities of having different risk grade of PCa by the presence of comorbid conditions were calculated using the values from Hjälm-Eriksson et al. 28 The probability of upgrading in PCa averaged 26.9% for Gleason score 3 + 3 and 22.6% for Gleason score 3 + 4, and the range of the variable was derived from the ranges of the 2 Gleason score levels as well.
Values in the range column were collected from the probabilistic sensitivity analyses results.

Sensitivity Analyses
We conducted 1-way sensitivity analyses on the test characteristics of the MRI and potential MRI-guided biopsy.Holding the sensitivity of MRI-guided biopsy constant, QALYs for the MRI and potential MRI-guided biopsy strategy decreased at very high sensitivity values of MRI.No additional incremental QALYs were gained for the MRI and potential MRI-guided biopsy strategy, with sensitivity values for the MRI greater than 0.84 across all PSA strata compared with 0.76 in the base case.
There were approximately 0.062 QALYs gained on average throughout the entire range of specificity values of MRI for the PSA stratum greater than 10 ng/mL, while the cost of the MRI and potential MRI-guided biopsy strategy decreased as the specificity of the MRI increased, resulting in an ICER less than $11 000 per QALY for the entire range.For the PSA stratum of 4.1 to 10.0 ng/mL, the ICER stayed below $50 000 per QALY for the entire range as the specificity increased.For the PSA stratum of 2.5 to 4.0 ng/mL, the specificity of the MRI needed to be at least 0.50 for the ICER to be less than $50 000 per QALY.
Holding the specificity of the MRI constant, the ICER remained at approximately $6000 per QALY or less for the entire range of specificity values of MRI-guided biopsy for the PSA stratum of more than 10 ng/mL.For the PSA stratum of 4.1 to 10.0 ng/mL and the PSA stratum of 2.5 to 4.0 Results from probabilistic sensitivity analysis indicate that the MRI and potential MRI-guided biopsy strategy was cost-effective at the $100 000-per-QALY threshold in a range between 76% and 81% of simulations for each of the 3 PSA strata of 2.5 ng/mL or more (Figure).Throughout the range of cost-effectiveness thresholds, the probability that the MRI and potential MRI-guided biopsy was cost-effective compared with the standard biopsy was found to be greater for higher PSA strata.

Discussion
This economic evaluation examined the cost-effectiveness of integrating annual MRI and potential MRI-guided biopsy compared with standard biopsy differentiated by PSA strata after PCa screening in the Medicare-eligible cohort.Results from this study suggest that the MRI and potential MRI-guided biopsy strategy is cost-effective compared with standard biopsy for the detection of PCa from a federal payer perspective among 65-year-old men with a PSA level of 2.5 ng/mL or more at  generally accepted WTP levels.Although for these men the initial cost of the MRI and potential MRI-guided biopsy strategy was higher than the cost of standard biopsy, extra costs were offset by higher expected QALYs.Our results are consistent with previous literature on the cost-effectiveness of MRI and potential MRI-guided biopsy strategy as management tools for PCa, 35 in addition to its already proven value in decreasing cancer overdiagnosis demonstrated in 2 prospective randomized clinical trials. 7,40Findings from our study are supported by 2 prior economic evaluations that compared MRI-guided biopsy with alternative strategies, including multiparametric MRI, which also found that MRI-guided biopsy is more cost-effective than the standard of care using TRUS biopsy. 41,42wever, PSA levels were not differentiated in those analyses.Barnett et al 35 examined the costeffectiveness of PCa screening strategies for people with PSA levels above 4 ng/mL and found that MRI and a combined biopsy in which both a standard biopsy and targeted MRI and ultrasonography fusion biopsy are performed is the optimal strategy, with an ICER of $23 483 per QALY.Pahwa et al 5 found that a standard biopsy is dominated by the MRI and cognitive-guided biopsy with an ICER of -$8946 per QALY and by the MRI and in-gantry MRI-guided biopsy with an ICER of -$1263 per QALY.
Our study stratifies the economic evaluation model by PSA, allowing us to examine MRI and potential MRI-guided biopsy outcomes in the context of this common marker, and we report ICERs using QALYs whose interpretation is broadly applicable.However, 1-way sensitivity analyses demonstrated no incremental QALY gain for MRI and potential MRI-guided biopsy for MRI sensitivity values greater than 0.84.This finding can be explained by the potential greater quality-of-life losses associated with early treatment compared with the risks and benefits associated with later treatment.We also account for both comorbidities and different grades of cancer in combination with PSA ranges.To our knowledge, our study is also the first economic evaluation that considered an annual MRI surveillance strategy.Prior studies have assessed the cost-effectiveness of MRI for PCa based on a single instance of screening, not accounting for costs of repeat MRI for patients receiving surveillance that occurs in clinical practice.Also, with increasing surveillance being associated with more frequent detections at lower PSA levels, especially in the 2.5 to 4.0 ng/mL range, there is a concern about the costeffectiveness of MRI due to the lower likelihood of cancer.However, even for men in this PSA stratum, our results indicated economic value gains for stakeholders.
A concern could be raised that cancers detected by MRI strategies could be more indolent than those detected by TRUS biopsy alone, and thus stratification data from TRUS biopsy could overestimate survival benefit from an MRI-guided strategy.However, the PROMIS (Prostate Magnetic Resonance Imaging Study) trial has previously shown that cancers missed by MRI strategies are smaller and lower in grade than those detected and, thus, the opposite is expected to be true. 41

Limitations
As is the case in all simulation studies, this study has a number of limitations.First, in estimating treatment cost, we applied equal proportions of the 3 most common treatment options for PCa.
Although changes in the proportion of treatment options will be associated with absolute cost estimates, varying the proportion of treatment will not be associated with the comparative results of the economic evaluation because any variation in cost or effectiveness (ie, treatment-related utility and survival) due to treatment proportions would be associated with both strategies equally.
Furthermore, treatment may vary and change over time (and by region) depending on clinician and patient preferences, 43 and incorporating such trends in economic models is important for research focused on budgetary effects.Future cost-effectiveness analyses of MRI and potential MRI-guided biopsy to screen for PCa should consider emerging treatment options and develop region-specific models that can more accurately project total cost and effectiveness outcomes to better inform local policy makers.
Developing a model necessitates making choices that may affect results and limit generalizability.We chose to use the cost for in-bore MRI-targeted biopsy instead of systematic TRUS fusion biopsy, as the former would detect the higher-risk grade PCa sooner, resulting in an expected increase in short-term costs and potentially decreasing the likelihood of cost-effectiveness.However,  44 Furthermore, our study assumes that men will stay in their original PSA stratum, which may not occur naturally with aging.However, because our results provide economic and effectiveness outcomes for each PSA category, concerns regarding dynamic changes in PSA can be addressed by weighting the results of PSA categories to calculate expected values for a model that incorporated those possible variations.We expect that including changes in PSA strata over time would not result in a qualitatively different conclusion because in our analysis the MRI and potential MRI-guided biopsy strategy are optimal in all groups with a PSA level of 2.5 ng/mL or more. 45r findings represent the widening of the PSA threshold from previous research by Shakir et al, 46 which found that MRI-guided biopsy was optimal compared with standard biopsy in detecting clinically significant PCa among men with a PSA level of more than 5.2 ng/mL.In addition, the likelihood of clinically significant cancer increases with increasing serum PSA value, although this is a multifactorial phenomenon, as prostatic hyperplasia is more likely to result in an elevated PSA, and prostatic hyperplasia would decrease the sensitivity of standard biopsy.For simplicity, this scenario was not incorporated into the model.

JAMA Network Open | Oncology
Our study was restricted to the Medicare-eligible cohort starting at 65 years of age for men followed up for 10 years after PSA screening.This cohort is particularly relevant for policy makers not only because of their increased risk of developing PCa but also because economic modeling tends to devalue later-life outcomes (ie, discounting), and thus the evaluation of surveillance strategies for PCa will undervalue cost reductions and QALY gains that occur during the age at which the disease burden is highest.Our choice of beginning the simulation at 65 years of age minimized the impact of such time discounting by focusing on the group with the highest incidence rate.Nevertheless, further investigation may be appropriate for younger men, using a model where initial screening costs would be incurred by other payers, and the association with quality of life might extend beyond 10 years.
Adjusting for the follow-up frequency, such as every other year rather than annual testing, could also be explored for individuals at low risk of PCa.
We did not further stratify our model by other patient characteristics such as family history, ethnic origin, diet, or genetic factors.This would require a markedly different model from the one used here as well as the availability of detailed longitudinal information on such cohorts.Finally, our study assumed a biopsy-naive population at enrollment.Further study is needed to determine optimal screening options for patients with a prior history of PCa or prior negative biopsy results, which would substantially alter the likelihood of receiving a diagnosis of clinically significant cancer.

Conclusions
This economic evaluation of a hypothetical cohort from a federal payer perspective suggests that investments in expanding the capacity to conduct MRI and potential MRI-guided biopsy will provide economic value for stakeholders and improve quality of life for 65-year-old men with a PSA of 2.5 ng/mL or more undergoing screening.Results from this study add to a growing consensus that the use of MRI and potential MRI-guided biopsy is cost-effective compared with standard strategies.
These findings are valuable as they can support establishing a new approach for men in active surveillance for PCa.The capacity for all men in the US to receive MRI and potential MRI-guided biopsy has been expanding, but greater adoption of this strategy is contingent on MRI availability and adequate resources and training for the MRI-guided procedures, as well as overcoming fixed cost barriers. 47

JAMA Network Open | Oncology
Annual Prostate MRI and MRI-Guided Biopsy for Men With Positive PSA Results

Table 1 .
PCa and Test-Related Probability Inputs a a The probability of PCa recurrence for different PSA strata is derived from Xia et al,

Table 2 .
Cost and Health Utility Weight Inputs a , the ICER stayed below $13 000 per QALY and $22 000 per QALY, respectively.For the PSA stratum of less than 2.5 ng/mL, the ICER decreased as the specificity increased, but it remained between $187 000 and $192 000 per QALY.Additional sensitivity analyses (eg, tornado diagrams) are provided in eFigure 2A and B in Supplement 1.
b Annual health care cost was based on a median of 3 years of total costs incurred for patients with prostate cancer and comorbidities.cHealthcarecost after treatment is an annual cost.JAMA Network Open | OncologyAnnual Prostate MRI and MRI-Guided Biopsy for Men With Positive PSA Results JAMA Network Open.2023;6(11):e2344856. doi:10.1001/jamanetworkopen.2023.44856(Reprinted) November 29, 2023 5/11 Downloaded from jamanetwork.comby guest on 12/14/2023 ng/mL

Table 3 .
Cost-Effectiveness Results a Incremental values and ICERs are in reference to standard strategy.
Annual Prostate MRI and MRI-Guided Biopsy for Men With Positive PSA Results these costs may be offset by potential long-term cost reductions due to greater efficacy of treatment when PCa is detected at an early stage, which may be of particular interest to policy makers.Future studies can address the difference in the estimated economic value of these different techniques once reliable longitudinal data are available.In addition, our outcomes may vary for other reimbursement settings and perspectives.A societal perspective, for example, could result in higher time costs for the patient or other costs not covered by Medicare that may differ between the strategies.