According to diagnosis year (A) and facility type (B) in adults and in younger adults according to diagnosis year (C) and facility type (D). All differences between adult vs younger adult groups according to number of lymph nodes examined significant at P < .001.
aP < .05 vs 2010.
bP < .05 vs 2011.
cP < .05 vs academic/research program.
dP < .05 vs comprehensive community cancer program.
eFigure. Study Cohorts for Analyses of Utilization of MMR Deficiency Testing in adult CRC Patients and Underuse of MMR Deficiency Testing in Young Adult CRC Patients
eTable. Coding of MMR Deficiency Testing According to Patient, Facility, Tumor, and Treatment Characteristics for All Adult Patients and Younger Adults with CRC
Customize your JAMA Network experience by selecting one or more topics from the list below.
Shaikh T, Handorf EA, Meyer JE, Hall MJ, Esnaola NF. Mismatch Repair Deficiency Testing in Patients With Colorectal Cancer and Nonadherence to Testing Guidelines in Young Adults. JAMA Oncol. 2018;4(2):e173580. doi:10.1001/jamaoncol.2017.3580
What are the predictors of utilization of mismatch repair deficiency testing in adults with colorectal cancer and nonadherence to testing guidelines in high-risk young adults?
In this study of 152 993 adults and, of these, 17 218 younger adult patients with colorectal cancer, only 28.2% and 43.1% underwent mismatch repair deficiency testing, respectively. Although the proportion of patients tested increased between 2010 and 2012, various sociodemographic, facility, tumor, and treatment factors were independently associated with underutilization and underuse.
Mismatch repair deficiency testing in patients with colorectal cancer remains low, even in high-risk populations; interventions tailored to groups at higher risk for nonadherence to testing guidelines may be warranted.
Mismatch repair (MMR) deficiency of DNA has been observed in up to 15% of sporadic colorectal cancers (CRCs) and is a characteristic feature of Lynch syndrome, which has a higher incidence in young adults (age, <50 years) with CRC. Mismatch repair deficiency can be due to germline mutations or epigenetic inactivation, affects prognosis and response to systemic therapy, and results in unrepaired repetitive DNA sequences, which increases the risk of multiple malignant tumors.
To evaluate the utilization of MMR deficiency testing in adults with CRC and analyze nonadherence to long-standing testing guidelines in younger adults using a contemporary national data set to help identify potential risk factors for nonadherence to newly implemented universal testing guidelines.
Design, Setting, and Participants
Adult (age, <30 to ≥70 years) and, of these, younger adult (<30 to 49 years) patients with invasive colorectal adenocarcinoma diagnosed between 2010 and 2012 and known MMR deficiency testing status were identified using the National Cancer Database. The study was conducted from March 16, 2016, to March 1, 2017.
Patient sociodemographic, facility, tumor, and treatment characteristics.
Main Outcomes and Measures
The primary outcome of interest was receipt of MMR deficiency testing. Multivariable logistic regression was used to identify independent predictors of testing in adult and/or young adult patients.
A total of 152 993 adults with CRC were included in the study (78 579 [51.4%] men; mean [SD] age, 66.9 [13.9] years). Of these patients, only 43 143 (28.2%) underwent MMR deficiency testing; the proportion of patients tested increased between 2010 and 2012 (22.3% vs 33.1%; P<.001). Among 17 218 younger adult patients with CRC, only 7422 (43.1%) underwent MMR deficiency testing; the proportion tested increased between 2010 and 2012 (36.1% vs 48.0%; P < .001). Irrespective of age, higher educational level (OR, 1.38; 95% CI, 1.15-1.66), later diagnosis year (OR, 1.81; 95% CI, 1.65-1.98), early stage disease (OR, 1.24; 95% CI, 1.18-1.30), and number of regional lymph nodes examined (≥12) (OR, 1.44; 95% CI, 1.34-1.55) were independently associated with MMR deficiency testing, whereas older age (OR, 0.31; 95% CI, 0.26-0.37); Medicare (OR, 0.89; 95% CI, 0.84-0.95), Medicaid (OR, 0.83; 95% CI, 0.73-0.93), or uninsured (OR, 0.78; 95% CI, 0.66-0.92) status; nonacademic vs academic/research facility type (OR, 0.44; 95% CI, 0.34-0.56); rectosigmoid or rectal tumor location (OR, 0.76; 95% CI, 0.68-0.86); unknown grade (OR, 0.61; 95% CI, 0.53-0.69); and nonreceipt of definitive surgery (OR, 0.33; 95% CI, 0.30-0.37) were associated with underuse of MMR deficiency testing.
Conclusions and Relevance
Despite recent endorsement of universal use of MMR deficiency testing in patients with CRC and well-established guidelines aimed at high-risk populations, overall utilization of testing is poor and significant underuse of testing among young adults persists. Interventions tailored to groups at risk for nonadherence to guidelines may be warranted in the current era of universal testing.
Colorectal cancer (CRC) is the third most common cancer in the United States and the second and third most common cause of cancer death in men and women, respectively.1 Mismatch repair (MMR) deficiency of DNA is a characteristic feature of Lynch syndrome and has been observed in up to 15% of sporadic CRCs. Mismatch repair deficiency can be due to germline mutations in DNA mismatch repair genes (MLH1, MSH2, MSH6, or PMS2) or somatic epigenetic silencing of MLH1, which result in unrepaired repetitive DNA sequences. These altered sequences increase the risk of multiple cancers, but are most commonly associated with CRC.
Historically, MMR deficiency testing was not uniformly recommended for all patients with CRC, but rather for individuals at higher risk for MMR deficiency based on personal or family history. The Amsterdam criteria (later updated to the Amsterdam II criteria)2 and the Bethesda guidelines (later updated to the revised Bethesda guidelines) were developed to help identify patients at high risk for Lynch syndrome according to age, personal cancer history, and family cancer history due to a greater risk of high microsatellite instability (MSI-H) tumors in these patients.3
Younger patients with CRC have an increased incidence of Lynch syndrome and MSI-H tumors. Due to this increased risk, long-standing national guidelines have recommended routine MMR deficiency testing for all patients with CRC younger than 50 years.3-5 Unfortunately, testing guidelines are not always well implemented, often resulting in underascertainment of patients with MSI-H tumors.6,7 Universal MMR deficiency testing was proposed by Hampel et al7 in 2008, but was not incorporated into the National Comprehensive Cancer Network guidelines until 2014.5 Currently, most national consortiums recommend universal screening for MSI-H tumors in all patients with newly diagnosed CRC.8-10 The purpose of this study was to evaluate recent utilization of MMR deficiency testing in adults with CRC using a large, contemporary, national data set during a period of increasing endorsement of universal testing. We also analyzed underuse of testing in patients with CRC younger than 50 years—a population with a well-established risk of MSI-H tumors—to identify and compare independent risk factors for potential nonadherence to recently implemented, universal testing guidelines.
The National Cancer Database (NCDB) is a program of the American College of Surgeons Commission on Cancer and the American Cancer Society. It is one of the largest clinical registries in the world and serves as a resource for clinical surveillance and quality improvement for cancer programs participating in the Commission on Cancer.11 The NCDB compiles data for more than 70% of all newly diagnosed cancer cases in the United States.12 Treatment data include the first course of therapy. This study was reviewed and approved by the Fox Chase Cancer Center Institutional Review Board. No patient consent was required for data analysis of an existing deidentified data set.
We identified all adult (age, <30 to ≥70 years) patients with CRC diagnosed between 2010 and 2012 in the NCDB participant user file. The study was conducted from March 16, 2016, to March 1, 2017. As shown in the eFigure in the Supplement, patients with CRC diagnosed before 2010 were excluded (due to low rates of reporting of MMR deficiency testing in the NCDB participant user file before this year), as well as patients without histologic diagnosis confirmation, noninvasive cancer, non–adenocarcinoma tumor histology, unknown tumor stage, or diagnosis at a reporting facility with treatment or decision not to treat done elsewhere (to optimize completeness and accuracy of data captured). To analyze underuse of MMR deficiency testing (ie, nonadherence to testing guidelines) in younger adult patients, we further limited our cohort to younger adult patients (age, 18 to 49 years).
The primary outcome of interest was receipt of MMR deficiency testing via immunohistochemistry (IHC) and/or polymerase chain reaction–based microsatellite instability analysis from initial diagnosis through the first course of treatment. Cancer registrars at participating NCDB sites were advised to utilize pathology reports, laboratory reports (internal or external), admission notes, or consultation notes (internal or external) to identify whether MMR deficiency testing was performed. Specific results of IHC or mutational status tests were not collected. Only patients with known MMR deficiency testing status were included in the present analysis (eFigure in the Supplement).
Charlson/Deyo comorbidity index was calculated by the International Classification of Disease, Ninth Revision, Clinical Modification secondary diagnosis codes (excluding cancer diagnoses and postoperative complications) obtained by registrars from discharge abstracts and billing records.13 Educational level and income were estimated at the zip code level based on percentage of adults not graduating from high school and median household income. Patient residence was determined by matching the state and county Federal Information Processing Standard code at the time of diagnosis. Cancer programs (denoted in the NCDB as reporting facilities) were categorized according to type and geographic location.
We used the Sequence Number field in the NCDB participant user file to identify patients who presented with a first cancer or had a previous cancer diagnosis. Tumor stage was categorized utilizing the Analytic Stage Group field, which assigns the value of the reported pathologic stage group or clinical stage group. We categorized tumor location as proximal (ie, cecum to transverse colon), distal (ie, splenic flexure to sigmoid colon), or rectosigmoid or rectal. Treatment characteristics evaluated included receipt of definitive surgical resection, type of resection, number of regional lymph nodes (LNs) examined (ie, <12 vs ≥12) in patients who underwent colectomy, and receipt of chemotherapy.
The objectives of this study were to evaluate utilization of MMR deficiency testing (ie, identify predictors of receipt of testing) in adult patients with CRC and underuse of recommended MMR deficiency testing in young CRC patients. Univariate associations between patient, tumor, facility, and treatment characteristics and MMR deficiency testing status and results were analyzed using χ2 tests or Cochran-Armitage trend tests. Multivariable logistic regression was used to identify independent predictors of receipt of MMR deficiency testing in adult (≥50 years) patients, as well as independent predictors of underuse of testing in younger adult (18-49 years) patients. All patient-, tumor-, facility-, and treatment-level covariates were included in the primary and secondary multivariate analyses. To account for the clustering within facility, we used robust SEs via generalized estimating equations with a working independence assumption.14 We used the Hosmer-Lemeshow test to evaluate goodness of fit. P < .05 was considered statistically significant, and all tests were 2-sided. Analyses were performed using SPSS statistical software, version 22 (SPSS Inc) and SAS software, version 9 (SAS Institute Inc).
Among 152 993 adults with CRC, only 43 143 (28.2%) were documented as undergoing MMR deficiency testing. A total of 78 579 (51.4%) men were included; mean (SD) age of the cohort was 66.9 (13.9) years. There was an increase in the proportion of patients tested between 2010 and 2012 (22.3% vs 33.1%; P < .001) (Table 1). Several patient sociodemographics, tumor characteristics, facility type and location, and treatment were associated with MMR deficiency testing status on univariate analysis (Table 1).
After controlling for all other covariates, female sex, higher educational level, later diagnosis year, previous cancer diagnosis, stage II disease, 12 or more regional LNs examined, and receipt of chemotherapy were independently associated with MMR deficiency testing (Table 2). In contrast, older age, African American race, lack of insurance or Medicare/Medicaid coverage, nonacademic or research facility type, stage 4 disease, rectosigmoid or rectal tumor location, histology (non–mucin-producing adenocarcinoma), unknown or lower tumor grade, and nonreceipt of definitive surgery or colectomy were independently associated with nonreceipt of MMR deficiency testing (Table 2).
Among 17 218 patients younger than 50 years, only 7422 (43.1%) were documented as undergoing MMR deficiency testing; as in the larger cohort of all adult patients, the proportion of patients tested increased between 2010 (36%) and 2012 (48%; P < .001) (Table 3). Several patient sociodemographics, tumor characteristics, facility type and location, and treatment were associated with MMR deficiency testing status on univariate analysis (Table 3).
After controlling for all other covariates, higher educational level, later diagnosis year, stage II disease, and 12 or more regional LNs examined were independently associated with MMR deficiency testing (Table 4). In contrast, older age (40-49 years), Hispanic ethnicity, lack of insurance or Medicare/Medicaid coverage, residence in a nonmetropolitan county, nonacademic or research facility type, facility location, rectosigmoid or rectal tumor location, unknown grade, and nonreceipt of definitive surgery were independently associated with nonreceipt of MMR deficiency testing (Table 4).
There was a strong association between number of regional LNs examined at colectomy and receipt of MMR deficiency testing both in the larger adult cohort and in younger patients. On multivariate analysis, examination of 12 or more regional LNs was the strongest predictor of receipt of MMR deficiency testing other than diagnosis year. We also assessed the association between the number of regional LNs examined and receipt of MMR deficiency testing according to diagnosis year and facility type (Figure). Within each diagnosis year, receipt of MMR deficiency testing was higher in patients who had 12 or more regional LNs examined, and rates of receipt of testing increased during the study period. Receipt of MMR deficiency testing was also higher in patients who had 12 or more regional LNs examined, irrespective of type of facility at which they were treated, although rates of receipt of testing were lower in nonacademic/research programs.
To our knowledge, this is the largest study using a contemporary, multi-institutional cohort to examine utilization and underuse of MMR deficiency testing in patients with CRC in the United States. Despite growing interest in and endorsement of universal testing in patients with CRC during the study period and long-standing, national guidelines recommending routine testing in all patients with CRC younger than 50 years, only 28.2% of adult patients and 43.1% of younger adults with CRC in our study cohorts underwent testing.3,7 Although the proportions of patients tested increased during the study period, our results suggest that underutilization of MMR deficiency testing was significant and pervasive, even among young patients with CRC with a well-established risk of Lynch syndrome. Our study provides important, national baseline data regarding utilization of MMR deficiency testing in patients with CRC and identifies significant groups at risk for potential nonadherence to newly implemented universal testing guidelines moving forward.
For nearly 3 decades, Lynch syndrome screening has been largely based on patient age and/or personal or family history of cancer.3-5 There are limited data regarding utilization of MMR deficiency testing in young patients with CRC, and it largely consists of significantly smaller, multi-institutional series. Kessels et al15 examined adherence to recommended MMR deficiency testing in 335 young patients treated for CRC between 1999 and 2005 at 1 of 7 academic or general hospitals in the Netherlands. As in our study, they found low rates of guideline adherence across the entire study cohort and in patients with stage IV disease. A similar study by Van Lier et al6 included 169 patients diagnosed with CRC who met inclusion criteria derived from the revised Bethesda guidelines and were treated at 11 Dutch hospitals between 2005 and 2006. Despite the fact that 10 of the hospitals referred their patients to the larger academic hospital for genetic counseling and MSI analysis, only 23% of eligible patients underwent MMR deficiency testing. Cross et al16 examined utilization of MMR deficiency testing in 1188 patients diagnosed with metastatic CRC between 2004 and 2009 at 7 integrated health care delivery organizations in the Cancer Research Network in the United States. Despite the availability of Lynch syndrome screening programs at these institutions, only 22.3% of 138 patients younger than 50 years underwent MMR deficiency testing. A population-based analysis by Karlitz et al17 of the Louisiana Tumor Registry found that only 23% of patients 50 years or younger underwent appropriate IHC/MSI screening. As in the present analysis, they noted several socioeconomic risk factors for nonadherence to guideline-based care. These earlier studies suggest that underuse of MMR deficiency testing exists across academic and nonacademic institutions, even in settings where access to genetic testing is adequate and well established. Our analysis, which used a large, more recent, national data set, supports and significantly extends these findings and suggests that underuse of MMR deficiency testing in the United States is a persistent and pervasive problem, even in clinical scenarios (ie, young adults with CRC) where its use was unequivocally recommended.
In addition to prognostic implications, the results of MMR deficiency testing may also have potential therapeutic implications. Although patients with MSI-H tumors typically have a better prognosis compared with patients with MSI low or stable tumors, they may not benefit from single-agent, fluorouracil-based adjuvant chemotherapy.18,19 Furthermore, there is a growing body of evidence suggesting that immunotherapy involving blockade of the programmed death-1 pathway can have a profound benefit in patients with MMR-deficient tumors. The US Food and Drug Administration recently granted accelerated approval to pembrolizumab for treatment of MSI-H or MMR-deficient solid tumors, and future trials in the adjuvant setting are planned.20 As such, MMR deficiency testing will likely have an increasingly important role in the care of patients with CRC.21
Several studies have evaluated the sensitivity of selective MMR deficiency testing and assessed utilization of universal screening in patients with newly diagnosed CRC.8-10,15,16 In a seminal study by Hampel et al,7 MMR deficiency testing was performed on 500 tumors from unselected patients with CRC. Among the 18 patients who ultimately tested positive for Lynch syndrome, only 44% were diagnosed before age 50 years and only 72% otherwise met the revised Bethesda criteria for testing. The authors concluded that selective testing would potentially fail to identify 28% of the cases of Lynch syndrome and instead proposed a strategy of active universal testing in patients with CRC. In a subsequent analysis by Julié et al22 of 2014 patients with newly diagnosed CRC, adoption of a universal testing strategy also identified more MSI-H tumors than a selective testing strategy based on revised Bethesda guidelines. A survey by Beamer et al23 demonstrated that, although 71% of National Cancer Institute Comprehensive Cancer Centers (which are classified as academic/research program facilities by the NCDB) universally reflexively screened for Lynch syndrome, only 15% of community cancer programs used universal testing. Our results demonstrate that, despite long-standing, national guidelines recommending MMR deficiency testing in young patients with CRC, there was significant and persistent underuse of testing during the study period. Although more recent studies suggest that universal screening in the United States is likely to be cost-effective, the present study suggests that sociodemographic characteristics and insurance status may prove to be substantial barriers to effectively implementing such a strategy across diverse patient populations and facilities.24
We identified various socioeconomic factors that were independently associated with underutilization and underuse of MMR deficiency testing. Although underutilization of MMR deficiency testing in older patients may reflect health care professionals’ perceptions about decreased risk of MSI-H with increasing age, underuse of testing according to socioeconomic and insurance status potentially suggests that national guidelines are not being equitably applied across all sociodemographic groups. Our results are in line with those of previous studies demonstrating disparities in receipt of guideline-based care in patients with CRC, particularly among nonwhite, non–privately insured, nonurban patients who tend to present with advanced disease and are less likely to receive optimal definitive therapy.11,25,26
Our study further demonstrated significantly lower rates of MMR deficiency testing across nonacademic/teaching cancer programs. Furthermore, we observed a strong association between the number of regional LNs examined and rates of MMR deficiency testing even after controlling for diagnosis year and facility type. More specifically, among patients who underwent colectomy, those who had 12 or more regional LNs examined (a consensus standard for hospital performance with colectomy for colon cancer) were significantly more likely to undergo MMR deficiency testing than were those who had fewer LNs examined.5,27 This association remained consistent across years of diagnosis, despite the fact that the proportion of patients who underwent MMR deficiency testing increased, potentially reflecting gradual adoption of universal screening guidelines during the study period. Although a similar association between adequacy of number of regional LNs examined and MMR deficiency testing was observed across facility types, lower rates of testing were observed at comprehensive community and community cancer/other programs. These findings are consistent with those of other studies demonstrating differences in guideline-based care according to type of treatment facility and suggest that efforts to improve rates of MMR deficiency testing among patients with CRC should ideally be embedded within more comprehensive, multidisciplinary efforts to continually optimize oncologic processes of care and outcomes at these sites.23,28-30
The present study has several potential limitations. The NCDB is not population-based, which may limit the generalizability of our findings.31 The available registry data may include misclassification due to coding errors and incomplete patient and treatment information. Although there were minor differences between our study cohorts and corresponding patients with unknown MMR deficiency testing (eTable in the Supplement), our findings may not be generalizable to these patients. Cancer registrars at Commission on Cancer sites that report to the NCDB track patients throughout the first course of treatment; however, it is possible that some patients underwent MMR deficiency testing at non–Commission on Cancer sites or after that time period. We were unable to determine whether the observed rates of underuse of MMR deficiency testing were partly due to underreferral. Finally, the study period predated the adoption of universal testing by several national consortiums; the current rates of testing may be significantly higher. The low rate of underuse of testing among young CRC patients noted in this study, however, demonstrates that guideline-based recommendations take a significant period of time to permeate routine care.
Our study demonstrates that, despite growing interest in and endorsement of universal testing for MMR deficiency in patients with CRC during the study period, as well as long-standing, well-established guidelines aimed at high-risk populations, utilization of testing in this population was limited and significant underuse of testing among younger adults persisted. Of particular concern, our study noted significant independent associations between patient socioeconomic and insurance status as well as cancer program location and type and utilization of MMR deficiency testing, suggesting that patient and health care system-level interventions tailored to groups at risk for nonadherence are warranted to ensure optimal and uniform implementation of newly endorsed, universal testing guidelines. Dissemination and implementation research are needed to enhance awareness and knowledge about Lynch syndrome among patients and health care professionals, increase referrals for genetic testing, and ensure adherence to both MMR deficiency testing guidelines and resulting cancer screening recommendations (eg, esophagogastroduodenoscopy, transvaginal ultrasonography).6,32-34 Analyses similar to the present study will be needed on an ongoing basis to track incremental progress toward closing this important clinical service gap.
Corresponding Author: Nestor F. Esnaola, MD, MPH, MBA, Department of Surgical Oncology, Fox Chase Cancer Center, 333 Cottman Ave, Philadelphia, PA 19111 (email@example.com).
Accepted for Publication: August 17, 2017.
Published Online: November 9, 2017. doi:10.1001/jamaoncol.2017.3580
Author Contributions: Drs Handorf and Esnaola have full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Shaikh, Meyer, Esnaola.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Shaikh, Esnaola.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Shaikh, Handorf, Esnaola.
Administrative, technical, or material support: Shaikh.
Study supervision: Meyer, Esnaola.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was supported by grant P30CA006927 from the National Institutes of Health, National Cancer Institute (Dr Handorf).
Role of the Funder/Sponsor: The National Institutes of Health, National Cancer Institute had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Create a personal account or sign in to: