Availability and utilization of molecular testing for primary central nervous system tumors among US hospitals

Abstract

Advanced molecular testing has increasingly become an integral component for accurate diagnosis of central nervous system (CNS) tumors. We sought to establish the current state of molecular testing availability and approaches for the diagnosis of CNS tumors in US hospitals that conduct high volumes of CNS tumor resections. We distributed a 16-item survey inquiring about molecular testing approaches for CNS tumors to 115 neuropathologists at US hospitals with neurosurgery residency programs. Thirty-five neuropathologists (30.4%) responded to the survey, all of whom indicated their institutions perform molecular testing on CNS tumor tissue. The most commonly offered tests were MGMT methylation profiling and next-generation sequencing. Fourteen respondents (40%) indicated that their institution is able to test for and report all of the molecular alterations included in our survey. Nine (25.7%) respondents indicated that molecular testing is performed as standard of care for all patients with resected CNS tumors. Our results suggest that even in academic hospitals with a high volume of CNS tumor resections, molecular testing for these tumors is limited. Continued initiatives are necessary to expand the availability of molecular testing for CNS tumors to ensure diagnostic accuracy and guide targeted therapy.

INTRODUCTION

The 2021 World Health Organization (WHO) Classification of Tumors of the Central Nervous System (WHO CNS5) is the second iteration in which identification of specific molecular alterations is necessary for diagnostic classification of central nervous system (CNS) tumors (1). Previously, tumor classification was based largely on histologic features, an approach limited by high interobserver variability (2). With the updated tumor classification system found in the fourth and fifth editions, the diagnostic criteria of tumors shifted from a reliance on morphological tumor characteristics in isolation to an emphasis on genetic alterations and biomarkers identified through molecular testing (3).

This new nosological classification of tumors, in both adult and pediatric populations, is expected to provide advantages in tumor treatment. For gliomas, ependymal tumors, and embryonal tumors, molecular-driven classification has increased accuracy in diagnosing, treating, and predicting their prognoses (4, 5). Among CNS tumors for which prognostic molecular alterations are limited, the detection of characteristic molecular alterations or methylation profiles can aid in the diagnosis (6). Targeting treatments to known genetic alterations found in tumors may lead to improved outcomes in some brain tumors (7–9). Recent positive clinical trials of IDH-, BRAF-, and NTRK-targeted therapy in gliomas have highlighted the importance of identifying putatively actionable alterations (10–12).

Although guidelines from the WHO do not recommend any specific method to derive this molecular information, the Consortium to Inform Molecular and Practical Approaches to CNS Tumour Taxonomy-Not Official WHO (cIMPACT-NOW) was created in 2017 to provide recommendations for the practical work-up of CNS tumors, primarily by neuropathologists (13). A total of 7 updates have been published by the cIMPACT-NOW to guide the integrated histological and molecular appraisal of CNS tumors (14–20). However, there is limited literature on the availability of molecular testing for CNS tumor characterization in US hospitals.

In 2021, Hopkins et al (21) reported that the average case volume of tumor resections among hospitals with neurosurgery residency programs was 607–628 cases per year. With a high volume of CNS tumor resections, we hypothesized that these institutions would routinely test for many of the molecular alterations necessary for accurately characterizing CNS tumors, according to the WHO CNS5. Thus, we sought to assess the availability of advanced molecular testing of CNS tumors by surveying neuropathologists at US hospitals with neurosurgery residency programs. We also investigated associations between genetic testing availability and the socioeconomic characteristics of the communities treated by these institutions.

MATERIALS AND METHODS

Survey design

Using the Qualtrics online survey platform, Version XM (Qualtrics, Provo, UT), we developed an electronic survey assessing molecular testing approaches for CNS tumors in US hospitals with neurosurgery residency programs. The survey consisted of 16 questions regarding molecular testing approaches for CNS tumors (Supplementary Data Fig. S1). Six questions asked if respondents’ institutions utilized the following tests to molecularly characterize CNS tumors and if they were performed internally or sent to another healthcare institution or laboratory: (1) next-generation sequencing (NGS) for DNA mutations; (2) RNA sequencing for fusions; (3) microarray-based comparative genomic hybridization; (4) fluorescence in situ hybridization (FISH); (5) O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation; and (6) DNA methylation profiling. In one question, participants were asked to check a box for each molecular alteration they were able to report in clinical care. We selected 21 characteristic molecular alterations included in the WHO CNS5, the vast majority of which are recommended by consensus guidelines (22–24). In the remaining questions, respondents were asked to report who is able to request molecular testing of CNS tumors, when these tests are performed in clinical care, who pays for the molecular testing, average turnaround times, and the average number of tumors that undergo molecular testing per month at their institution. The names of the respondents’ associated institutions were collected. Otherwise, survey responses were anonymously collected and no identifying information about respondents was collected.

Survey distribution and study population

We identified 115 neurosurgical residency training programs from the American Association of Neurological Surgeons website (Supplementary Data Table S1). Neuropathologists from these institutions were identified from institutional websites. We selected neuropathologists with the title of “associate professor” or “professor,” with either a clinical or research focus on oncology. If more than one neuropathologist at the institution had an oncology focus, we chose the neuropathologist with the longest tenure at the institution. Pathology residents and fellows were excluded. We also reviewed institutional websites to determine whether the institution had a molecular diagnostics lab, molecular pathology division, or a molecular fellowship. From February 2023 to September 2023, we emailed all eligible neuropathologists a letter describing the purpose and confidentiality of the study, as well as a hyperlink to complete the Qualtrics survey. Surveys were distributed in 4 phases. Initially, we reached out to the neuropathologist with the most extended tenure at the institution, with an oncological focus. For those who did not respond initially, we followed up with a second attempt. In the third phase, we contacted another neuropathologist with an oncological focus from institutions lacking a respondent. Finally, in the fourth phase, we recontacted all neuropathologists who had not responded. Survey responses were stored in Qualtrics. Survey responses with >50% incomplete or missing data were excluded from the final analysis. This study was approved by our Institutional Review Board (IRB00361329).

Assessment of municipal socioeconomic characteristics

Publicly available demographic and economic data for the cities in which respondents’ institutions are located were obtained from the US Census Bureau website (data.census.gov). The following continuous variables were collected from the 2022 US Census Bureau website: population size, percent Hispanic, percent White, percent Black, percent foreign-born, percent of the population with a bachelor’s degree or higher, percent of the population without health insurance, and median household income.

Statistical analysis

Survey answers were summarized using standard descriptive statistics. Statistical analyses were conducted using IBM SPSS Statistics Software (Version 29.0.0; IBM Corp., Armonk, NY). Wilcoxon-rank sum tests were used to compare continuous variables. Spearman’s rank correlation tests were used to evaluate correlations between the number of molecular tests available at each institution and municipal socioeconomic characteristics. Two-tailed p-values <.05 were used as the limit of statistical significance.

RESULTS

We requested participation from 115 neuropathologists from 115 US hospitals with neurosurgery residency programs and received 38 total survey responses. Among the total responses received, 3 were excluded on the basis of incomplete responses, yielding 35 complete responses to be included in the final analysis. The overall response rate was 30.4%, which mirrors a previously reported response rate for a web-based survey distributed to physician specialists (25). The geographic distribution of respondents’ institutions is presented in Figure 1.

Figure 1.

Geographic distribution of respondents’ associated institutions.

All respondents indicated that their institution offers molecular testing for CNS tumor tissue. Three (8.6%) institutions perform all molecular testing internally, 5 (14.3%) institutions send all samples to outside facilities for molecular testing, and the remaining 27 (77.1%) institutions utilize a combination of internal and external testing platforms (Fig. 2). Among the 5 institutions that send all samples to outside facilities for molecular testing, 3 had molecular pathology divisions or fellowships, according to their websites. DNA methylation profiling was most commonly performed as a send-out test (82.9%), while FISH was most commonly performed internally (62.9%). Send-out locations included reference laboratories at other academic hospitals (40.0%) and commercial laboratories (51.4%). The most common send-out institutions included the National Institutes of Health, Mayo Clinic, Caris Life Sciences, and Tempus.

Figure 2.

Molecular testing approaches among survey respondents’ institutions.

Respondents were asked whether their institution or affiliated outside hospital/laboratories were able to test and report alterations in the following genes—IDH1, IDH2, BRAF, EGFR, TERT promoter, TP53, ATRX, FGFR1/2/3, CKDN2A, histone H3 genes, PTEN, MYB, NF2, PRKCA, YAP1, MN1, and ZFTA. The methods for assessing copy number changes in chromosome 7, chromosome 10, chromosome 1p, and chromosome 19q, as well as the method for assessing MGMT promoter hypermethylation, were also assessed. The results are shown in Table 1. The number of institutions that were able to test for a given molecular alteration in our survey was significantly correlated with the year the molecular alteration was included in the WHO Classification of CNS Tumors. Molecular alterations that were included in earlier editions of the WHO Classification of CNS Tumors were more commonly tested (r = −0.602, p = .004). All respondents reported that their institution or affiliated send-out location tested for 1p/19q codeletion, IDH1, and IDH2 mutations, which were all included in the 2016 WHO Classification of CNS Tumors. The least commonly tested mutations were in the ZFTA and MN1 genes, which were not recommended until 2021. Only 14 respondents (40.0%) indicated that their institution tests all of the genes included on the survey.

Table 1.

Frequency of institutions that perform advanced molecular testing of the following targets

Gene	N	%
IDH-1 mutation	35	100.0
IDH-2 mutation	35	100.0
1p/19q codeletion	35	100.0
BRAF mutations and structural alterations/fusions	33	94.3
EGFR amplification	33	94.3
MGMT methylation	32	91.4
TERT promoter mutation	32	91.4
TP53 mutation	31	88.6
ATRX mutation	30	85.7
FGFR1, FGFR2, FGFR3 mutations and structural alterations/fusions	29	82.9
CDKN2A homozygous deletion	29	82.9
Histone H3 gene mutations (including K27 and G34)	28	80.0
PTEN mutations and homozygous deletion	28	80.0
MYB structural alterations/fusions	26	74.3
NF2 mutations and structural alterations/fusions	26	74.3
Trisomy of chromosome 7, monosomy of chromosome 10	26	74.3
PRKCA mutations and structural alterations/fusions	26	74.3
MYBL1 structural alterations/fusions	24	68.6
YAP1 structural alterations/fusions	22	62.9
MN1 structural alterations/fusions	21	60.0
ZFTA structural alterations/fusions	20	57.1

We also assessed molecular testing protocols. Nine (25.7%) institutions performed molecular testing as standard of care for all patients with resected CNS tumors. Twenty-four (68.6%) institutions perform molecular testing upon pathologist request. For the remaining 2 institutions that do not perform advanced molecular testing as standard of care or upon pathologist request, advanced molecular testing is ordered by a neurooncologist. Furthermore, outside of pathology, many institutions accommodate requests for advanced molecular testing from various providers. Nineteen (54.3%) institutions perform molecular testing upon neurosurgeon request. Twenty-seven (77.1%) institutions perform molecular testing upon oncologist request. Seven (20.0%) institutions perform molecular testing upon patient request.

At 29 (80.0%) institutions included in the analysis, the patient or patient’s insurance pays for molecular testing for primary CNS tumors. Three (8.6%) respondents indicated that their associated institutions cover the costs of this test.

The median (range) number of tumors tested/month was 12.5 (2–100). The median turnaround time for test results is 14 days (IQR: 14–21). The number of tumors tested each month was significantly correlated with the number of molecular alterations tested. Institutions that performed more molecular testing on tumors each month offered a greater number of molecular tests (r = 0.426, p = .012). Notably, institutions that perform NGS internally had quicker turnaround times than institutions that send-out samples for this testing (mean ± SD: 13.1 ± 4.1 days vs 17.2 ± 5.0 days, p = .038). Institutions that perform NGS internally also tested more tumors per month than institutions that send-out samples for this testing (mean ± SD: 32.2 ± 27.2 vs 8.6 ± 6.1, p = .004).

We evaluated correlations between sociodemographic characteristics of the institutions’ city and the availability of molecular testing for primary brain tumors. Molecular testing availability was not associated with the racial or ethnic makeup of the city in which the institution is located. However, the number of molecular alterations an institution could test was positively correlated with municipal median household income (r = 0.536, p = .001).

DISCUSSION

With the publication of the fourth and fifth editions of the WHO Classification of CNS tumors, molecular profiling is now essential to the diagnosis and classification of many CNS tumors. The 5-year window between the publication of the 2016 revision of the fourth edition and the fifth edition was the shortest interval between editions to date, representing accelerated advances in molecular diagnostic paradigms and improved understanding of the molecular etiology of these tumors. As the guidelines for classifying CNS tumors rapidly evolve, it is important to consider the current state of molecular testing for CNS tumors in the institutions responsible for diagnosing and treating patients with CNS tumors. In this study, we aimed to address this gap in knowledge by surveying neuropathologists at US hospitals with neurosurgery residency programs, institutions that historically perform high volumes of CNS tumor resections (21). All respondents reported the use of molecular testing for surgically resected CNS tumors. However, only 40% of respondents reported that their associated institutions test and report all mutations included in our survey. Our findings suggest that the implementation of molecular profiling in routine clinical care may be lagging behind the rapid advances in genetic research and modern classification recommendations.

All respondents reported that their institutions have the capability to test for mutations in IDH1, IDH2, and chromosome 1p/19 codeletion, 3 genetic alterations necessary for subtyping adult-type diffuse gliomas per the WHO, which account for ∼80% of malignant CNS tumors (26). In addition to IDH1 and IDH2 mutations and chromosome 1p/19q codeletion, the WHO CNS5 also incorporates TERT promoter mutations, EGFR amplification, or gain of chromosome 7 and loss of chromosome 10 into the diagnostic molecular criteria of glioblastoma, IDH-wildtype. These alterations are not covered by all respondents, with 91.4% of respondents reporting that their institutions test for TERT promoter mutations, 94.3% testing for EGFR, and 74.3% testing for trisomy of chromosome 7 and monosomy of chromosome 10. Furthermore, in the WHO CNS5, the presence of CDKN2A homozygous deletion is used to upgrade tumors to Grade 4 in the case of “astrocytoma, IDH-mutant,” or Grade 3 in the case of “oligodendroglioma, IDH-mutant and 1p/19q-codeleted,” even in the absence of high-grade histologic features (mitotic figures, microvascular proliferation, necrosis) (16). Eighty-three percent of respondents indicated that their institutions test for CDKN2A homozygous deletion. Overall, these findings suggest that even in academic hospitals with high volumes of CNS tumor resections, nearly 20% of gliomas may not receive the level of diagnostic accuracy recommended by the WHO (27). Differentiation of pediatric-type diffuse gliomas is largely based on MAP kinase and H3 alterations. Eighty percent of respondents reported that their institution tests for H3 mutations. For rarer CNS tumors, such as glioneuronal tumors, supratentorial ependymomas, and astroblastoma, for which alterations in YAP1, MN1, and ZFTA are included in the diagnosis, accurate characterization is even more limited.

In addition to driving more reliable and precise diagnoses of CNS tumors, comprehensive molecular profiling of CNS tumors allows providers to devise more targeted therapeutic approaches. In a cross-sectional study of patients with various cancer types, Marquart et al (2) found that ∼6% of cancer patients benefit from “genome-informed” therapy. One of the most profound examples of this practice is the utilization of DNA methylation profiling to determine MGMT promoter methylation status and copy number, which are predictive of response to alkylating chemotherapy agents in patients with gliomas (28). Since this alteration is not crucial for diagnosing gliomas but rather serves to predict responsiveness to therapy, pathologists may not routinely order this test during the initial evaluation. Our data are limited because we did not specifically identify who requested this testing. Furthermore, BRAF p. V600E-mutant gliomas have been shown to be responsive to RAF inhibitors (dabrafenib), MEK inhibitors (trametinib), and BRAF inhibitors (vemurafenib) (24, 29, 30). Over 90% of institutions represented by our survey respondents evaluate MGMT methylation and BRAF alterations. Gliomas harboring fusions in NTRK and BRAF have also exhibited responsiveness to targeted therapy (31, 32). Eighty-eight percent of survey respondents indicated that their institutions perform RNA sequencing for fusions.

Although a high proportion of respondents’ institutions test for molecular alterations that may help guide treatment decision-making, only 25.7% of respondents reported that molecular testing is performed as standard of care for all patients with resected tumors. The relatively low percentage of patients undergoing advanced molecular testing may be attributed to variations in how the survey question was interpreted. This represents a notable limitation of our survey, which warrants attention in future studies. For instance, neuropathologists might opt for advanced molecular testing in cases where identifying characteristic molecular alterations is imperative for diagnosis but not necessarily for CNS tumors like meningiomas where diagnosis primarily relies on histopathological features alone. Although characteristic alterations are acknowledged for these tumors, their significance in prognostic stratification and treatment guidance remains incompletely understood, yet represents a focal point of ongoing research efforts (33–36). The majority of respondents do not perform molecular testing unless requested by the neurosurgeon, oncologist, or pathologist. Similarly, prior research suggests that advanced molecular testing is underutilized in the initial treatment decisions for patients with CNS tumors (37). Instead, targeted gene sequencing is mostly used to assess eligibility for clinical trials or to guide therapy following tumor recurrence (38). While we did not specifically survey neuropathologists on their specific reasons for requesting advanced molecular testing, our findings suggest that this responsibility is frequently shared between neurosurgeons, oncologists, and pathologists. Thus, molecular testing may guide multiple aspects of care, from the decision to perform a repeat resection to the selection of adjuvant therapy. A struggle often encountered by neurosurgeons, oncologists, pathologists, clinical laboratory staff, and other hospital personnel is the economic constraint of whether the cost of ordering and performing diagnostic tests is justified based on the clinical utility of the test (38). Future qualitative studies are warranted to understand the decision-making process driving neurosurgeons, neuropathologists, and neuro-oncologists to request this testing. These data may be used to guide education aimed at increasing providers’ understanding of the importance of integrating this testing into high-value clinical care.

In addition to providers’ understanding of the utility of these tests in the management of patients with CNS tumors, there are many barriers to accessing sequencing technology, including the cost, lack of coverage from insurance, underrepresentation of minorities, and availability of resources to accommodate the technology, including insufficient information technology support (39–41). While we did not find a correlation between the availability of molecular testing and municipal demographic characteristics, we observed an association between the number of molecular alterations tested at an institution and the median household income of the institution’s city. One study found that despite the clinical advantages NGS has for the treatment of cancer patients, there are reported racial and socioeconomic disparities present in access to the sequencing technique to identify tumor biomarkers (42). In 2020, Lamba et al demonstrated that patients who were elderly, uninsured, Medicaid insured, lower income, or from rural communities were less likely to receive MGMT testing for newly diagnosed glioblastoma than counterparts (43, 44). The authors noted that even with increasing incidence of this test across the United States, these patient populations continued to experience this disparity. We were constrained in assessing associations solely between municipal demographic characteristics and institutional advanced molecular testing capacity. We acknowledge that this may not capture patient-level associations comprehensively. Future institutional or population-based studies investigating patient demographic characteristics associated with more or less robust molecular testing for CNS tumors may be warranted. Interestingly, 20% of our survey respondents reported that molecular testing may be performed upon patient request. While prior evidence suggests that patients desire to play an active role in treatment decisions, it is likely that the knowledge to request this testing may be limited to patients with a high level of medical literacy, which may further introduce disparities in advanced molecular testing for CNS tumors (45). To mitigate these disparities, it is imperative that patients receive molecular testing for CNS tumors, regardless of their socioeconomic background. Our survey is limited in its exploration of the complexities inherent in paying for advanced molecular testing. However, our findings suggest that increased coverage by insurance companies may be necessary to support widespread molecular testing, given that 80% of respondents indicated patients’ insurance covered the cost of molecular testing.

In addition to patient socioeconomic factors, Lamba et al (43) found that type of hospital had an impact on MGMT testing, with patients treated at community hospitals receiving MGMT testing less frequently than patients treated at academic institutions. Given this prior evidence, and our results showing that only 40% of our survey respondents from academic institutions with robust neurosurgery departments tested all genes included on the survey, we hypothesize the reduced availability of certain tests may be more pronounced in nonacademic community hospital settings. Smaller institutions may also not have a large volume of neuropathologists to drive the decision to perform molecular testing. However, large population-based studies are necessary to test this hypothesis. A potential solution is increased collaboration between hospitals and diagnostic laboratories (46). Specifically, if smaller community institutions have reduced availability of diagnostic testing, partnering with neighboring academic institutions to perform more advanced tests may aid in improving the accuracy of diagnosing primary CNS tumors. The practice of partnering with other institutions was apparent in our cohort. Forty percent of represented institutions send samples to other hospitals for molecular testing and 51.4% send samples to outside laboratories. Approximately 32% of respondents reported the use of private companies for diagnostic testing. One limitation associated with outside testing is longer turnaround times, as indicated by our findings. With an increasing number of clinical trials being driven by biomarkers, expedited molecular testing is becoming essential. Reduced turnaround times are crucial to avoid missing the eligibility window for clinical trials.

We recognize several limitations in this study, including the small sample size and response bias inherent in many survey-based studies. Albeit these limitations, our study indicates that even in academic hospitals with neurosurgery residency programs, advanced molecular testing of primary CNS tumors is limited. In light of rapidly advancing molecular testing approaches and discovery in the field of neuro-oncology, continued initiatives are necessary to implement these advances in routine patient care.

FUNDING

This study received no direct funding and author contributions were done voluntarily.

CONFLICT OF INTEREST

The authors do not have any relevant conflicts of interest.

SUPPLEMENTARY DATA

Supplementary Data can be found at academic.oup.com/jnen.