Bioethical implications of current state practices of molecular diagnostics in neuropathology

Although advanced molecular diagnostic services have become integral in neuro-oncology,¹ effective health care policies for their deployment remain lacking. The WHO’s 2021 classification guidelines further merge molecular testing (mol-testing) for clinical decision making, which precipitates the need to find ethical, equitable, and logistically feasible workflows for molecular neuropathology.² The unfavorable insurance coverage of mol-testing represents a major barrier to effective deployment, and novel solutions have emerged which require bioethical assessment.

Insurance administration in the United States is markedly complex. Lack of FDA approval commonly results in advanced mol-testing claim denial. However, FDA approvals require time-consuming investments often unfeasible for local molecular laboratories.³ Prior ambiguity of FDA jurisdiction of laboratory-developed tests (LDTs), resulted in oversight by the Clinical Laboratory Improvement Amendment (CLIA) to ensure clinical laboratory safety/quality standards. While CLIA was thought to act in lieu of FDA approval, the 2018 VALID Act broadly encompassed LDTs as in vitro clinical tests (IVCTs) covered under FDA administration and CLIA certification.³ Although CLIA-certified laboratories may perform testing using non-FDA-approved IVCTs, associated costs from non-approved FDA tests can fall upon patients or hospitals. These fiscal realities have thus hindered mol-testing development and execution.

In this wake, private companies—such as Foundation, Caris, Tempus, and others—have emerged to fill the gap in coverage for advanced mol-testing. Testing costs thus move from the patient or hospital to the private company as a partnered mol-testing firm. However, the business models driving many of these firms are to provide services in exchange for sequencing data and deidentified patient records. These business models range from direct data monetization for contract research or offering cutting-edge, high-resolution patient data for research purposes to generate new contracts with hospitals. Unlike academic centers, patient data heavily drive the profits needed for firm sustainability and reinvestment toward R&D pipelines. While private firms meet requirements set forth by HIPAA to protect patient data, additional oversight through institutional review boards or ethical review committees may not be present given the deidentified nature of the data.⁴ Such internal oversight allows for independent evaluation of data management, data use, and scrutiny of intent. Thus, external review of these practices is critical to evaluate bioethical implications independent of firm interests.⁴

To frame the ethical conundrum, consider 1p/19q testing. Fluorescence in-situ hybridization (FISH) is easily reimbursable and commonly available yet suffers from a ~5% false-positive rate.⁵ Thus, in some cases, subsequent non-reimbursed testing (eg, NGS or chromosomal microarray) is required for confirmation. Reliance on private companies for testing services is done to mitigate costs in exchange for data. This model gives rise to potential patient exploitation, an ethical concern not often attended to in the literature yet highly relevant to neuro-oncology. On one prominent bioethical model, exploitation is understood as a kind of transactional unfairness. We can distinguish the unfairness in the transactional outcomes (one party receiving too much benefit or one party taking on too much burden) from unfairness in the transactional process (eg, lack of informed consent).⁶ On this model, it is the unfairness in the process that constitutes wrongful exploitation. Therefore, to evaluate whether such a model is overly exploitative, we must determine whether patients are fully consenting to a company’s terms by whether patients (1) understand the breadth of releasing their molecular data and (2) feel undue pressure to agree.

To the first point, while there is a financial value of genomic data, a patient may be inadequately informed and thus (i) not appreciate such value and (ii) not understand data is being used for commercial purposes.⁷ Patients are readily agreeable to releasing data when it improves diagnostic accuracy and/or the disease’s knowledge-base.⁸ However, data storage between firms and the ordering group limits data accessibility without purchase—which may hinder scientific innovation and conflict with patient values. Moreover, studies have demonstrated patient hesitation regarding the commercialization of their data without proper renumeration.^7,9

To the second point, brain tumors can be highly aggressive and require accurate classification—creating time pressure for physicians and patients. In consequence, firms may exploit these underlying pressures, forcing agreements even if doing so were to conflict with patient values. In this way, patients are in a vulnerable position regarding their choice to authorize the commercialization of their data. There are therefore important ethical questions as to whether patients fully understand the commercialization of their data or feel undue pressure to sign off on such commercialization—and whether it is within the neuro-oncologist’s duties to notify patients of these issues.

Notably, just because an agreement is exploitative, it is unclear whether third parties should intervene and block them. It can be argued that such a model harms neither the patient nor physician, but rather improves patient outcomes at the cost of placing financial gains in the hands of private, for-profit firms. It is possible that the underlying factors that drive the patient’s choices fit within an exploitative paradigm, but continuing this model is still more favorable than cessation. If regulatory bodies intervened to block such practices while not providing feasible alternatives, the resulting lower diagnostic accuracy would produce net harm for patients.⁵

Private firms may continue to play an important role given that most academic centers cannot meet this challenge without philanthropic backing. Although a subset of firms has succeeded in gaining FDA approval to perform reimbursable testing without complete reliance on data monetization strategies, they continue to utilize data by submitting hospitals as a key marketing strategy to gain contracts. As most firms remain pre-initial public offering, a need to evaluate the long-term business sustainability of these models is warranted. Ultimately, the role of private firms has arisen due to the complexity needed to generate reimbursable FDA-approved IVCTs. Organizations without the administrative capacity to overcome this hurdle experience insurance declines of their IVCTs. However, as molecular targets are recognized with therapies tailored to them, NGS and other testing (eg, methylation profiling) will be critical in establishing standards of care that improve disease stability and reduce hospitalizations. In turn, this may improve the cost-effectiveness of care that insurers ultimately want.¹⁰ We posit that insurance-based revenue streams for testing pose fewer bioethical concerns than data monetization approaches, so long as the reimbursement levels continue to provide incentives for private firms to invest in their R&D. In closing, external evaluation of private mol-testing firms in neuro-oncology is critical for ethical data use. Furthermore, improvements in the current FDA approval process and recognition of advanced mol-testing by insurers are needed to ultimately change the current practices in the field.

Acknowledgments

W.W. acknowledges support from NIH NCATS TL1TR002735 and UL1TR001450. J.J.O. and D.T. acknowledge support from NIH NINDS R03NS116334. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

References