Molecular Testing in Anatomic Pathology and Adherence to Guidelines: A College of American Pathologists Q-Probes Study of 2230 Testing Events Reported by 26 Institutions.
There are several chemotherapeutic regimens that rely upon ancillary testing to guide therapy. The National Comprehensive Cancer Network (NCCN) guidelines (1) are widely used and they provide recommendations for the use of ancillary testing in various tumor types. Guidelines for lung carcinoma, (2) colorectal carcinoma, (3) and melanoma (4) (current at the time this study was conducted ) are briefly summarized herein. For metastatic or locally advanced nonsquamous, non-small cell lung carcinoma, the NCCN guidelines include epidermal growth factor receptor gene (EGFR) mutation analysis and anaplastic lymphoma kinase gene (ALK) rearrangement testing by fluorescence in situ hybridization (FISH), with testing for ROS1 gene rearrangement by FISH listed as an additional consideration. For metastatic colorectal adenocarcinoma, KRAS and BRAF mutation testing appears on the NCCN guidelines, but EGFR mutation testing is not recommended. In metastatic melanoma, NCCN guidelines include BRAF mutation analysis, as the V600E or V600K mutations of the BRAF gene guide vemurafenib therapy. In addition, KIT-mutated melanomas may be treated with imatinib. A recently published guideline from the College of American Pathologists, International Association for the Study of Lung Cancer, and Association for Molecular Pathology (CAP/IASLC/AMP) specifically addresses EGFR and ALK testing for lung carcinomas. (5)
This Q-Probes study aimed to assess how frequently molecular testing in anatomic pathology adhered to NCCN guidelines and to determine if any practice or demographic characteristics were related to this rate. In addition, data were collected to determine testing turnaround times and adequacy of various specimen types for molecular testing.
MATERIALS AND METHODS Data Collection
The study consisted of 2 data collections. One was a retrospective collection limited to cases of lung primary carcinoma, colorectal primary carcinoma, and melanoma for which molecular testing was performed or attempted during the prior year, or until 40 cases were collected, whichever came first. The second was a prospective collection of all cases of solid tumors for which molecular testing was requested during a period of 2 months or until 30 cases were collected, whichever came first. For both data collections participants recorded the following for each case: primary tumor site; tumor histology; ordering provider; specimen type; extent of clinical disease; which molecular tests were requested, performed, or attempted; and the primary rationale for the test request. The study authors then compared these data points to the NCCN guidelines as were current at the time of data collection (beginning early 2013) and assessed each molecular test as either (1) strictly meeting the guideline, (2) loosely meeting the guideline, or (3) not meeting the guideline. In addition, turnaround time information and specimen adequacy were reported for the retrospective study. Both data collections excluded breast primary carcinomas, hematolymphoid tumors, and gynecologic cytology. Lynch syndrome screening was excluded from the retrospective study but included in the prospective study. Both data collections included in-house and send-out testing.
To classify the molecular testing events and to assess adherence to guidelines, the following definitions were applied.
Molecular Test.--Analysis for the presence or quantitation of nucleic acids (either DNA or RNA) and proteins with the aim to guide therapy, predict prognosis, diagnose disease, or predict occurrence of disease. Examples applied to anatomic pathology specimens include gene sequencing, DNA hybridization techniques (eg, fluorescence in situ hybridization, chromogenic in situ hybridization, microarrays), and amplification methods (eg, polymerase chain reaction).
Strict Adherence to Guidelines.--The test is included in the NCCN guideline as an integral part of the decision-making algorithm (eg, EGFR testing in metastatic or locally advanced lung adenocarcinoma, KRAS testing in metastatic colorectal adenocarcinoma).
Loose Adherence to Guidelines.--The test is not part of the NCCN decision-making algorithm but it falls into either of the following categories: (1) the test is directly related to other standard tests in the guideline, making its performance reasonable in certain situations (eg, KRAS or ROS1 testing in metastatic or locally advanced lung adenocarcinoma) or (2) the test is mentioned in the guideline as a consideration for certain patients (eg, Lynch syndrome screening in colon adenocarcinoma, BRAF testing in metastatic colorectal adenocarcinoma).
For both data collections the primary performance indicators were the percentage of tests that strictly met the NCCN guidelines and the percentage of tests that at least loosely met the guidelines. There were 2 additional performance indicators for the retrospective study of lung carcinoma, colon carcinoma, and melanoma: the percentage of cases with adequate tissue for molecular testing and the median turnaround time from molecular test request to test result.
Statistical analysis was performed to determine which factors were significantly associated with the performance indicators. Participant results were excluded from the analysis for those indicators based on fewer than 5 tests. Associations between the performance indicators with the demographic and practice variables were analyzed by using Kruskal-Wallis tests for discrete-valued independent variables and regression analysis for the continuous independent variables. Variables with significant associations (any P value < .10) identified by this initial analysis were then included in a forward selection multivariate regression model for which a significance level of .05 was used. The distributions of the aggregate test characteristics were compared by using the [chi square] tests and Fisher exact test; a significance level of .05 was used for these tests. All statistical tests were performed with SAS 9.2 (SAS Institute, Cary, North Carolina).
Twenty-six institutions submitted data for this Q-Probes study. Twenty-five (97%) were located in the United States and 1 in Brazil. Sixteen (62%) were teaching hospitals and 12 (46%) had pathology residency training programs.
Within the past 2 years, all 26 participating laboratories had been inspected by the CAP, and 3 (11%) had been inspected by The Joint Commission. Seventeen (65%) institutions were urban, 5 (19%) were suburban, and 4 (15%) were rural. Fifteen (58%) were voluntary nonprofit hospitals, 6 (23%) were nongovernmental university hospitals, and there was 1 laboratory from each of the following categories: governmental nonfederal university, other nongovernmental, proprietary hospital, system/integrated delivery network, and veterans hospital. Distribution of bed size was as follows: 0-150, 8 (31%); 151-300, 7 (27%); 301450, 2 (8%); 451-600, 5 (19%); and greater than 600, 4 (15%).
In the prior year (2012), participating laboratories had accessioned a median of 16 963 surgical pathology cases (10th to 90th percentile range, 4832-40 425) and a median of 2507 nongynecologic cytology cases (10th to 90th percentile range, 276-11 530).
Retrospective Study of Lung, Colon, and Melanoma
The retrospective portion of the study consisted of a collection of testing that had been either performed or attempted on cases of colon primary carcinoma, lung primary carcinoma, and melanoma. Twenty-six institutions reported data on 1508 testing events. In aggregate, 996 of 1508 (66%) strictly adhered to NCCN guidelines and 1319 of 1508 (87%) at least loosely met guidelines. For all institutions the median for strict adherence was 71% and median for loose adherence was 95% (Table 1). The median turnaround time was 8 days, with a 10th to 90th percentile range of 4 to 13 days. The aggregate turnaround time was wide, ranging from same day to 127 days. There was adequate tissue to obtain a test result in a median 98%, with a 10th to 90th percentile range of 86% to 100%. Details of the individual cases are listed in Table 2. Pathologist-initiated reflex testing accounted for 483 of 1508 of testing (32%), while medical oncologist orders accounted for 56%. The tested specimen was an excision in a minority of cases (498 of 1508, 33%), with the remainder of specimens consisting of smaller biopsy samples. The molecular tests performed or attempted are listed in Table 3; EGFR, ALK, and KRAS were the most common by a large margin. Details regarding adequacy of material by specimen type are in Table 4; cell blocks showed a significantly lower adequacy rate than other specimen types (P < .001). Lynch syndrome screening was to be excluded from the retrospective study, but some participants did include such cases (93 of 1508 test events, 6.2% of total).
Prospective Study of All Case Types
The prospective portion of the study consisted of a collection of requests for molecular testing regardless of organ site. Twenty-three institutions reported data on 722 testing events. In aggregate, 360 of 722 (50%) strictly adhered to NCCN guidelines and 645 of 722 (89%) at least loosely met guidelines. For all institutions the median for strict adherence was 53% and the median for loose adherence was 94% (Table 1). Details for the individual cases are listed in Table 5. The most common organ sites were the lower gastrointestinal tract and lung. Pathologist-initiated reflex testing accounted for 303 of 722 tests (42%) and medical oncologists accounted for 325 of 722 test requests (45%). Again, excision specimens were in the minority (303 of 722, 42%). Details regarding the molecular tests requested are listed in Table 6. The most common requests were for EGFR, KRAS, ALK, Lynch syndrome screening, and BRAF.
Participants also completed a general questionnaire regarding various aspects of institutional practices and handling of some specific testing situations (Tables 7 through 12). Most laboratories do not have a written policy for turnaround time goals in molecular testing. There was variation in how molecular test orders are handled, including 4 of 23 participants (17%) reporting that neither MD nor PhD personnel are involved in the process. Slightly more than half of participants require pathology department approval for some molecular testing, and generally such approval is required for send-out testing, proprietary tests, and testing for inpatients. One institution required pathology approval for all molecular tests. It appears that some effort is made to have block and slide selection performed by a pathologist familiar with the case, though a small minority of laboratories (2 of 23, 9%) have designated a specific pathologist for this function. In terms of creating an integrated report, 15 of 23 laboratories (65%) routinely incorporate molecular test results into anatomic pathology reports, and two-thirds of those participants routinely include correlative information as to the significance of the test results. Participation in tumor boards was common, with general tumor boards (18 of 23, 78%) and breast conferences (18 of 23, 78%) the most common. Nearly all of the participating institutions (19 of 21, 90%) hold cancer accreditation of some form. Additional practice details are listed in Table 7.
Pathologist-initiated reflex testing is performed in slightly more than half of participating laboratories (13 of 23, 57%), and most (8 of 13, 62%) of these institutions do not require a standing written order for such reflex testing. The most commonly performed reflex tests were immunohistochemistry for mismatch repair proteins (12 of 13, 92%), followed by ALK and EGFR testing on lung tumors with appropriate histology (8 of 13, 62% each) and KRAS testing on colon cancers (8 of 13, 62%). A minority of laboratories (5 of 13, 39%) use a sequential or algorithmic approach to reflex testing. Most laboratories (18 of 23, 78%) are able to meet the 10-day turnaround time goal for EGFR and ALK testing on lung cases. Additional details are listed in Table 8.
Participants reported details on special handling techniques for small biopsy samples and cytology specimens in anticipation of molecular testing. Roughly half of laboratories handle lung biopsy samples differently from other specimen types, including cutting additional unstained slides or making fewer initial levels for hematoxylin-eosin staining. Preliminary adequacy assessment for fine-needle aspiration specimens is provided in 21 of 23 laboratories (91%). Additional details are listed in Table 9.
Laboratories reported which molecular tests were performed in-house versus send-out in their practices (Table 10). Most tests are send-outs, and 10 laboratories do not perform any in-house molecular tests. For in-house testing, participants reported various details regarding micro-dissection and reporting. The Medicare 14-day rule was not consistently followed, as more than half of participants (6 of 11, 55%) disregard the rule for in-house testing and 9 of 22 (41%) disregard the rule for send-out testing. Additional details are listed in Table 11.
There was variation in the approach to screening for Lynch syndrome, with 6 of 23 (26%) using an algorithm involving reflex immunohistochemistry and polymerase chain reaction testing and an equal proportion routinely performing both tests concurrently. The remainder varied by clinician request and individual pathologist practice. Similarly, there was wide variation in how patients were selected for testing. Additional details are provided in Table 12.
Practice Characteristics and the Effect on the Performance Indicators
The performance indicators were tested for associations with institutional demographic and practice variables. A P value below .05 was considered statistically significant. Although no associations were identified between the performance indicators and the demographic or practice characteristics, 2 observations were found for the retrospective study of lung, colon, and melanoma cases. First, institutions holding a higher number of multidisciplinary conferences tended to have a higher rate of molecular testing at least loosely meeting guidelines (P = .07). Second, compared to colon cancer and melanoma, lung cancer testing more often strictly met guidelines (74% lung, 44% colon, 61% melanoma, P < .001) and more often at least loosely met guidelines (90% lung, 81% colon, 82% melanoma, P < .001). Two similar observations were found for the prospective study of all case types. First, testing on lung cancer more often strictly met guidelines when compared to other specimen sites, such as melanoma and lower gastrointestinal tract (71% lung, 63% melanoma, 29% colon, 39% aggregate other, P < .001). Second, testing on lung, colon, brain/spinal cord, and skin/melanoma more often at least loosely met guidelines when compared to all other sites (96% lung, 87% colon, 96% melanoma, 96% brain/spinal cord, 82% aggregate other, P < .001), based on the aggregate test results. Rural institutions had a longer molecular testing turnaround time than institutions located in city/suburban areas. This observation was not formally tested because of the low sample size (4 rural and 22 city/ suburban institutions), but the rural turnaround time was a median 13 days (range, 8-16 days), whereas city/suburban turnaround time was a median 7 days (range, 1-12 days).
This Q-Probes study examined the rate of adherence to NCCN guidelines for molecular testing, the turnaround time for molecular tests, and the adequacy of specimens for testing. The study included both in-house and send-out testing on anatomic pathology specimens. Excluded from the study were cases of hematolymphoid neoplasia, breast primary carcinomas, and gynecologic cytology specimens.
The study consisted of 2 data collections: 1 retrospective and 1 prospective. The retrospective collection consisted of testing that had been previously performed or attempted and was restricted to lung carcinoma, colorectal carcinoma, and melanoma. In addition to allowing a more focused assessment of solid tumors that are commonly tested, the retrospective collection enabled assessment of turnaround time and adequacy of material for test completion. The retrospective data collection revealed that a median 71% of cases strictly met guidelines and a median 95% at least loosely met guidelines. In the retrospective collection the most common tests were for EGFR, ALK, and KRAS, in aggregate comprising 74% (1111 of 1508) of tests.
The prospective data collection consisted of requests for molecular testing and included a broader spectrum of organ sites. This data collection enabled assessment of the breadth of molecular test requests, the distribution of tumor types and organ sites undergoing testing, and allowed capture of any test requests that, for various reasons, ultimately may not have been completed. For the prospective collection, a median of 53% of cases strictly met guidelines and a median 94% at least loosely met guidelines. The most common tests requested were for EGFR, KRAS, ALK, Lynch syndrome screening, and BRAF, in aggregate comprising 86% (621 of 722) of the test requests.
Ranges of guideline adherence were quite broad among the participating institutions. Institutions holding a higher number of multidisciplinary conferences tended to have a higher percentage of testing that met guidelines. Tests were most frequently requested and/or performed on lung and lower gastrointestinal tract specimens. Testing on lung carcinomas more often met NCCN guidelines than for other organ sites. Perhaps this is related to the higher profile and increased emphasis on molecular testing of lung cancers through both the NCCN and the recently published CAP/ IASLC/AMP guideline for EGFR and ALK testing. (5) Median turnaround time for molecular tests was 8 days, which meets the recommended 10-day turnaround time goal set by the CAP/IASLC/AMP guideline. From our study's general questionnaire, 78% of participants report they are able to routinely meet the 10-day turnaround time goal for EGFR and ALK testing in lung cancer. Thus, the 10-day goal seems reasonable. A sequential or algorithmic approach to lung testing has potential to reduce the number of tests performed, but only a minority (5 of 13, 39%) of our study participants reported such an approach, and this was likely because of turnaround time concerns. In a recent survey of National Cancer Institute-designated cancer centers, average turnaround time for lung cancer testing with a sequential approach was 22.8 days, as compared to 7.6 days when tests were performed concurrently. (6) Most laboratories (18 of 23, 78%) in our study do not have a formal written goal for turnaround time in molecular testing in anatomic pathology.
This study provides some insights into pathologist involvement in molecular testing and the important function of the pathologist as gatekeeper. In our study most specimens used for molecular testing were small biopsy samples, highlighting the pathologist's continuing challenge to accomplish ever more with ever less material. In aggregate, 96% (751 of 785) of specimens had adequate material for testing, but there were statistically significant differences for the specimen type, with a cytology cell block adequacy rate of only 84% (80 of 95). It is notable that 91% (21 of 23) of participants routinely perform preliminary adequacy assessments for fine-needle aspiration specimens, thus the cell block adequacy rate in our study may actually be higher than typical. There were too few data points to link specimen adequacy to molecular laboratory dissection practices. Although pathologist-initiated reflex testing comprised a minority of the testing in our study, 57% (13 of 23) of laboratories reported that they routinely provide some form of either reflex or algorithmic testing, and most of those laboratories (8 of 13, 62%) do not require a standing written order for the testing. Perhaps the other laboratories require written orders as a means to avoid any appearance of inappropriate self-referral. The most common reflex testing was for Lynch syndrome screening in colon carcinomas, followed by ALK and EGFR for lung carcinomas, and KRAS for colon carcinomas. Similar to other studies, (7-9) we found considerable variability in both the selection of patients for Lynch syndrome screening and in the performance of the various test components, highlighting the need for standardization in this area.
More than half of participants report that pathology department approval is required for at least a subset of molecular tests, with approval most often required for send-out testing, tests for inpatients, and proprietary tests. There was inconsistent consideration of the 14-day rule for Medicare patients; for in-house testing, 55% (6 of 11) of laboratories ignore the rule and perform testing at the time of request, but only 41% (9 of 22) ignore the rule for send-out testing. This likely reflects an attempt to balance financial concerns with turnaround time pressures, but laboratories must tread lightly since inadvertent submission of inappropriate charges to Medicare may risk false claims accusations. In most practices orders for molecular testing are handled by a pathologist or PhD scientist, but 17% (4 of 22) of laboratories reported that such orders are typically handled by non-MD, non-PhD personnel. Most laboratories (15 of 23, 65%) incorporate molecular testing into a comprehensive anatomic pathology report, with 67% (10 of 15) of those laboratories routinely adding correlative information regarding the significance of the molecular testing results.
There are limitations to the data we report. We chose NCCN guidelines because they are comprehensive and widely used, but we recognize that other treatment guidelines exist. The field of molecular pathology is rapidly evolving, thus the appropriateness of testing on specific tumors or organ sites is fluid and ongoing clinical trials will continue to alter practice patterns. The NCCN guidelines have been updated multiple times since the data collection for this study began. Assessing adherence to NCCN guidelines was not always clear cut and there were some gray areas involved, hence the designations of "strict adherence" and "loose adherence" to guidelines. In some cases the ambiguity was related to the guidelines themselves, in others it was due to limitations in our data collection, and in others it was due to limitations in what participants reported. One example of guideline ambiguity is screening for Lynch syndrome. While NCCN encourages testing of all patients with colorectal carcinoma who are younger than 50 years or patients with stage II disease, it does not state specific criteria for Lynch syndrome testing. Some centers have implemented screening of all colorectal and endometrial cancers, regardless of age at diagnosis or family history, and this approach has been endorsed for colorectal cancer by the Evaluation of Genomic Applications in Practice and Prevention Working Group from the Centers for Disease Control. (9) Similarly, ROS1 testing for lung cancer is a "consideration" in patients who are ALK negative. For melanoma, KIT mutation analysis is listed as part of "other active regimens" without strict guidance as to which patients should be tested.
To keep the data collection manageable for participants, the required data points for each submitted case were kept to a minimum and, in some instances, this affected assessment of guideline adherence. For example, we did not ask for tumor grade; O-6-methylguanine-DNA-methyltransferase gene (MGMT) testing is not appropriate for all gliomas but is restricted to high-grade tumors. We also did not delineate specific locations of head and neck squamous cell carcinomas; human papilloma virus testing is not necessarily appropriate for every site in the head and neck. We did not collect data regarding prior treatment regimens; KIT and platelet-derived growth factor receptor a gene (PDGFRa) mutation testing is not necessarily warranted for every patient with gastrointestinal stromal tumor. We also did not collect data regarding appropriateness of specific testing methodologies (eg, EGFR mutation analysis versus EGFR amplification by fluorescence in situ hybridization). Assessment of testing appropriateness is not possible for tumors of unknown primary site.
Limitations in participants' submitted data included some cases with clearly inaccurate data entered (eg, "bone primary adenocarcinoma") and some write-in responses for test names that were unrecognizable. For lung primary carcinomas, the study input form listed all possible carcinoma histologic subtypes, but some cases were submitted with histology of "other," making it likely that at least some of the data were reported by nonpathologists.
Our data collection did not ascertain if molecular testing was performed on all cases that could have potentially undergone testing, thus we do not have a denominator or any means of measuring underutilization of molecular testing. Our data also do not address what inappropriate testing may have been performed despite objection by the pathologist. Particularly in the community setting, where a pathologist's livelihood depends on providing good customer service, it can be difficult to deny insistent clinicians. We also do not know if the results of the molecular tests were acted upon, which could be the focus of a future study. Finally, considering our small sample size of 26 institutions, nearly all of which held cancer center accreditation, our data are likely not generalizable to all of pathology practice.
In summary, this Q-Probes study aimed to assess the proportion of anatomic pathology molecular testing that adheres to NCCN guidelines. We are not aware of any prior multi-institutional study of this topic. The study is limited by a low number of participants, but it does provide useful benchmarking data and important information regarding how laboratories are handling this rapidly growing facet of pathology practice.
The authors thank the participating institutions for their time and diligence in collecting the data for this study.
(1.) National Comprehensive Cancer Network. NCCN clinical practice guidelines in oncology. 2013. http://www.nccn.org/professionals/physician_gls/f_ guidelines.asp#site. Accessed March 30, 2013.
(2.) National Comprehensive Cancer Network. NCCN clinical practice guidelines in oncology--non small cell lung carcinoma. Version 2. 2013. http://www nccn.org/professionals/physician_gls/pdf/nscl.pdf. Accessed March 30, 2013.
(3.) National Comprehensive Cancer Network. NCCN clinical practice guidelines in oncology-colon cancer. Version 3. 2013. http://www.nccn.org/ professionals/physician_gls/pdf/colon.pdf. Accessed March 30, 2013.
(4.) National Comprehensive Cancer Network. NCCN clinical practice guidelines in oncology--melanoma. Version 2. 2013. http://www.nccn.org/ professionals/physician_gls/pdf/melanoma.pdf. Accessed March 30, 2013.
(5.) Lindeman NI, Cagle PT, Beasley MB, et al. Molecular testing guideline for selection of lung cancer patients for EGFR and ALK tyrosine kinase inhibitors: guideline from the College of American Pathologists, International Association for the Study of Lung Cancer, and Association for Molecular Pathology. Arch Pathol Lab Med. 2013; 137(6):828-1174.
(6.) Schink JC, Trosman JR, Weldon, CB, et al. Biomarker testing methods in breast, gastric, and lung cancers: a benchmarking survey of NCI cancer centers. / Clin Oncol. 2013; 31(suppl):Abstract e22093.
(7.) Beamer LC, Grant ML, Espenschied CR, et al. Reflex immunohistochemistry and microsatellite instability testing of colorectal tumors for Lynch syndrome among US cancer programs and follow-up of abnormal results. J Clin Oncol. 2012; 30:1058-1063.
(8.) Bellcross CA, Bedrosian SR, Daniels E, et al. Implementing screening for Lynch syndrome among patients with newly diagnosed colorectal cancer: summary of a public health/clinical collaborative meeting. Genet Med. 2012; 14(1):152-162.
(9.) Evaluation of Genomic Applications in Practice and Prevention (EGAPP) Working Group. Recommendations from the EGAPP Working Group: genetic testing strategies in newly diagnosed individuals with colorectal cancer aimed at reducing morbidity and mortality from Lynch syndrome in relatives. Genet Med. 2009; 11(1):35-41.
Keith E. Volmar, MD; Michael O. Idowu, MD, MPH; Rhona J. Souers, MS; Raouf E. Nakhleh, MD
Accepted for publication November 24, 2014.
From the Department of Pathology, Rex Pathology Associates, Raleigh, North Carolina (Dr Volmar); the Department of Pathology, Virginia Commonwealth University, Richmond (Dr Idowu);the Department of Biostatistics, College of American Pathologists, Northfield, Illinois (Ms Souers); and the Department of Pathology, Mayo Clinic Jacksonville, Jacksonville, Florida (Dr Nakhleh).
The authors have no relevant financial interest in the products or companies described in this article.
Reprints: Keith E. Volmar, MD, Rex Pathology Associates, 4420 Lake Boone Trail, Raleigh, NC 27607 (e-mail: keith.volmar@ rexhealth.com).
Table 1. Molecular Testing Adherence to National Comprehensive Cancer Network Guidelines All Institutions Percentiles n 10th 25th Median 75th 90th Retrospective study (lung, colorectal, melanoma) Percentage of tests 26 32.6 64.7 70.9 82.7 89.7 that strictly meet the guideline Percentage of tests 26 57.4 90.7 95.1 98.9 100.0 that at least loosely meet the guideline Prospective study (all case types) Percentage of tests 23 20.0 31.4 53.3 66.7 70.5 that strictly meet the guideline Percentage of tests 23 75.0 87.0 94.3 100.0 100.0 that at least loosely meet the guideline Table 2. Details of Retrospective Study (Lung, Colon, Melanoma) No. Percentage Case type Lung carcinoma 499 60.7 Colon carcinoma 251 30.5 Melanoma 72 8.8 Histology Adenocarcinoma (invasive) 598 73.4 Melanoma 73 9.0 Non-small cell carcinoma, not otherwise specified 49 6.0 Carcinoma, not otherwise specified 26 3.2 Squamous cell carcinoma 20 2.5 Adenocarcinoma in situ 15 1.8 Adenosquamous carcinoma 5 0.6 Large cell neuroendocrine carcinoma 5 0.6 Large cell carcinoma, not otherwise specified 3 0.4 Small cell carcinoma 3 0.4 Carcinoid/low-grade neuroendocrine tumor 1 0.1 Other 17 2.1 Ordering provider Medical oncologist 451 56.4 Pathologist (reflex test) 257 32.1 Surgeon 61 7.6 Radiology oncologist 6 0.8 Generalist 3 0.4 Other 22 2.8 Specimen type Excision 268 32.6 Core biopsy 186 22.6 Incisional or open biopsy 163 19.8 Cytology cell block 96 11.7 Cytology smear 5 0.6 Other small biopsies (eg, forceps biopsies) 104 12.7 Extent of clinical disease Metastasis proven 354 43.1 Not known 247 30.0 Locally confined 120 14.6 Metastasis suspected 51 6.2 Locally advanced 44 5.4 Recurrence or refractory 6 0.7 Primary reason for test request Predictive/guide therapy 515 62.8 Not known 138 16.8 Diagnostic 91 11.1 Prognostic 74 9.0 Patient request 1 0.1 Research/clinical trial 1 0.1 Adequate material for testing Yes 752 95.7 No 34 4.3 No. of molecular tests performed or attempted per case 1 387 47.4 2 261 32.0 3 110 13.5 4 27 3.3 5 25 3.1 6 6 0.7 Table 3. Tests Performed or Attempted in Retrospective Study (Lung, Colon, Melanoma) Molecular Test Performed or Attempted No. Percentage EGFR 446 29.6 ALK (EML4-ALK) 358 23.7 KRAS 307 20.4 BRAF 114 7.6 Lynch syndrome screening (MMR/MSI) 93 6.2 NRAS 39 2.6 PIK3CA 39 2.6 ROS1 24 1.6 KIT 22 1.5 HER2/ERBB2 (only nonbreast cases) 5 0.3 ERCC1 3 0.2 HRAS 2 0.1 MET 2 0.1 Other 54 3.6 Other molecular test, submitted text response Proprietary tests (45) ... ... Unrecognized tests (9) ... ... Abbreviations: ALK, anaplastic lymphoma receptor tyrosine kinase (anaplastic lymphoma kinase);BRAF, B-Raf proto-oncogene, serine-threonine kinase (v-raf murine sarcoma viral oncogene homolog B); EGFR, epidermal growth factor receptor;EML4, echinoderm microtu-bule-associated protein-like 4; ERCC1, excision repair cross-comple-mentation group 1; HER2-ERBB2, v-erb-b2 avian erythroblastic leukemia viral oncogene homolog 2 (human epidermal growth factor receptor 2); HRAS, Harvey rat sarcoma viral oncogene homolog;KIT, v-kit Hardy-Zuckerman 4 feline sarcoma viral oncogene homolog;KRAS, Kirsten rat sarcoma viral oncogene homolog; MET, MET proto-oncogene, receptor tyrosine kinase;MMR, mismatch repair protein; MSI, microsatellite instability; NRAS, neuroblastoma rat sarcoma viral (v-ras) oncogene homolog; PIK3CA, phosphatidylinositol-4,5-bisphos-phate 3-kinase, catalytic subunit a; ROS1, ROS protooncogene 1, receptor tyrosine kinase. Table 4. Specimen Type and Adequacy for Molecular Testing (Retrospective Study of Lung, Colon, Melanoma) No. of No. With Adequate Specimen Type (a) Specimens Material (%) Excision 251 250 (99.6) Core biopsy 176 163 (92.6) Incisional or open biopsy 155 152 (98.1) Cytology cell block 95 80 (84.2) Cytology smear 5 4 (80.0) Other small biopsies (eg, forceps biopsies) 103 102 (99.0) (a) Statistically significant difference between cytology cell block and all other types (Fisher exact test;P < .001). Molecular Testing in Anatomic Pathology-Volmar et al 1118 Table 5. Details of Prospective Study (All Case Types) No. Percentage Tumor site Lower gastrointestinal tract (small bowel, colon) 177 35.8 Lung 171 34.6 Skin 24 4.9 Brain and spinal cord 19 3.8 Genital, female 19 3.8 Lymph node 14 2.8 Upper gastrointestinal tract (esophagus, stomach, duodenum) 14 2.8 Thyroid 11 2.2 Liver and biliary tract 9 1.8 Nasopharynx-oropharynx- hypopharynx 7 1.4 Unknown primary 7 1.4 Soft tissue 6 1.2 Kidney 5 1.0 Genital, male 2 0.4 Breast (exclude primary carcinoma) 1 0.2 Salivary gland 1 0.2 Other 7 1.4 Histology Adenocarcinoma (invasive) 359 72.7 Melanoma 30 6.1 Carcinoma, not otherwise specified 27 5.5 Glioma 18 3.6 Non-small cell carcinoma, not otherwise specified 16 3.2 Squamous cell carcinoma 14 2.8 Adenocarcinoma in situ 5 1.0 Large cell neuroendocrine carcinoma 4 0.8 Sarcoma 3 0.6 Gastrointestinal stromal tumor 1 0.2 Large cell carcinoma, not otherwise specified 1 0.2 Small cell carcinoma 1 0.2 Other 15 3.0 Ordering provider Medical oncologist 216 45.0 Pathologist (reflex test) 201 41.9 Surgeon 50 10.4 Radiology oncologist 7 1.5 Geneticist 2 0.4 Other 4 0.8 Specimen type Excision 208 42.3 Core biopsy 82 16.7 Incisional or open biopsy 76 15.4 Cytology cell block 47 9.6 Other small biopsies (eg, forceps biopsies) 79 16.1 Extent of clinical disease Metastasis proven 202 40.9 Not known 138 27.9 Locally confined 90 18.2 Locally advanced 36 7.3 Metastasis suspected 23 4.7 Recurrence or refractory 5 1.0 Primary reason for test request Predictive/guide therapy 285 57.9 Prognostic 117 23.8 Diagnostic 70 14.2 Not known 19 3.9 Research/clinical trial 1 0.2 No. of requested molecular tests per case 1 254 55.8 2 142 31.2 3 52 11.4 4 7 1.5 Table 6. Test Requests From Prospective Study (All Case Types) Molecular Test Requested No. Percentage EGFR 163 22.6 KRAS 159 22.0 ALK (EML4-ALK) 114 15.8 Lynch syndrome screening (MMR/MSI) 113 15.7 BRAF 74 10.2 HER2/ERBB2 (only nonbreast cases) 14 1.9 MGMT methylation 14 1.9 K/T 13 1.8 HPV in situ hybridization (excluding GYN cytology) 9 1.2 ROSI 8 1.1 1p and 19q deletion 6 0.8 /DH1 and/or /DH2 mutation 4 0.6 NRAS 2 0.3 PAX8-PPARA 2 0.3 HPV genotype (excluding GYN cytology) 1 0.1 HRAS 1 0.1 MET 1 0.1 PDGFRA 1 0.1 Other 23 3.2 Other molecular test, submitted text response Proprietary tests (19) ... ... MSH6 (1) ... ... TP53 (1) ... ... Thymidylate synthase (1) ... ... Unrecognized test (1) ... ... Abbreviations: ALK, anaplastic lymphoma receptor tyrosine kinase (anaplastic lymphoma kinase);BRAF, B-Raf proto-oncogene, serine-threonine kinase (v-raf murine sarcoma viral oncogene homolog B); EGFR, epidermal growth factor receptor;EML4, echinoderm microtu-bule-associated protein-like 4;GYN, gynecologic; HER2-ERBB2, v-erb-b2 avian erythroblastic leukemia viral oncogene homolog 2 (human epidermal growth factor receptor 2);HPV, human papilloma virus; HRAS, Harvey rat sarcoma viral oncogene homolog;-DH, isocitrate dehydrogenase;K-T, v-kit Hardy-Zuckerman 4 feline sarcoma viral oncogene homolog; KRAS, Kirsten rat sarcoma viral oncogene homolog;MET, MET protooncogene, receptor tyrosine kinase (protooncogene-C-Met); MGMT, O-6-methylguanine-DNA-methyltransfer-ase;MMR, mismatch repair protein;MSH6, mutS homolog 6;MSI, microsatellite instability; NRAS, neuroblastoma rat sarcoma viral oncogene homolog;PAX8-PPARA, paired box 8-peroxisome prolifer-ator-activated receptor a; PDGFRA, platelet-derived growth factor receptor, a polypeptide;ROSI, ROS proto-oncogene 1, receptor tyrosine kinase (c-ros oncogene 1); TP53, tumor protein p53. Table 7. Laboratory Practices Related to Management and Reporting of Molecular Tests No. Percentage Does your laboratory have a written policy for turnaround time goals for molecular testing on solid tumors in anatomic pathology? Yes 5 21.7 No 18 78.3 Who typically manages orders for molecular testing on solid tumors in anatomic pathology in your department? Any available pathologist 10 43.5 Non-MD, non-PhD personnel 4 17.4 Select pathologist(s) specifically 4 17.4 designated for this function Other 4 17.4 Other category, submitted text response ... ... Case pathologist (2) Combination of MD, non-MD, ... ... non-PhD personnel (1) PhD specifically assigned ... ... for case (1) N/A, such orders are typically 1 4.3 managed through standing protocols Is pathology department approval required for any molecular tests on solid tumors in anatomic pathology in your institution? Yes 12 52.2 No 11 47.8 If yes, which of the following require pathology department approval? (a) (n = 12) Send-out molecular tests 11 91.7 Molecular tests on inpatients 8 66.7 Proprietary molecular tests 7 58.3 Molecular tests above a cost threshold 1 8.3 Other 1 8.3 Other category submitted text response All molecular testing (1) Who typically chooses the blocks or slides used for molecular testing on solid tumors in anatomic pathology? Pathologist familiar with the case 14 60.9 Any available pathologist 7 30.4 Select pathologist(s) specifically 2 8.7 designated for this function Do you incorporate the results of molecular testing for solid tumors into anatomic pathology reports? Yes, either an addendum or amendment is 15 65.2 issued to the anatomic pathology report No 8 34.8 If yes, do you routinely provide correlative information regarding the significance of the molecular test results? Yes 10 66.7 No 5 33.3 Which multidisciplinary conferences are held at your institution? (a) (n = 23) Breast 18 78.3 General tumor board 18 78.3 Gastrointestinal 11 47.8 Genitourinary 11 47.8 Gynecologic 11 47.8 General surgical conference 9 39.1 Hematopathology 9 39.1 Infectious disease 9 39.1 Liver 9 39.1 Neurosurgery 9 39.1 Renal 9 39.1 Thoracic 9 39.1 Endocrine 8 34.8 Head and neck 7 30.4 None 1 4.3 Does your institution hold accreditation as a recognized cancer program? (a) (n = 21) Yes, American College of Surgeons 15 71.4 Commission on Cancer Yes, other cancer program accreditation 5 23.8 No, my institution is not a cancer center 2 9.5 Abbreviation: N/A, not applicable. (a) Multiple responses allowed. Table 8. Laboratory Practices Related to Reflex Testing and Turnaround Time No. Percentage Does your laboratory routinely initiate reflex molecular testing on solid tumors in anatomic pathology? Yes 13 56.5 No, we do not perform reflex 10 43.5 molecular testing If yes, do you require a standing order for reflex testing? Yes 5 38.5 No 8 61.5 If yes, what reflex molecular testing do you routinely perform on solid tumors in anatomic pathology? (a) (n = 13) MMR immunohistochemistry for colon cancers 12 92.3 ALK (EML4-ALK) for lung cancers with 8 61.5 appropriate histology EGFR for lung cancers with 8 61.5 appropriate histology KRAS for colon cancers with 8 61.5 appropriate histology KRAS for lung cancers with 7 53.8 appropriate histology MSI polymerase chain reaction 7 53.8 for colon cancers BRAF for melanomas 5 38.5 BRAF for colon cancers with 3 23.1 appropriate histology ROS1 for lung cancers with 1 7.7 appropriate histology If yes, does any of your reflex testing on solid tumors follow a sequential or algorithmic approach? Yes 5 38.5 No 8 61.5 What is your most common approach for EGFR and EML4-ALK testing in non-small cell lung carcinomas? Both tests are performed at the same 15 65.2 time (only EGFR and EML4-ALK are performed) These 2 tests are performed as part 7 30.4 of a broader test panel A sequential/algorithmic approach is 1 4.3 used (involving only EGFR and EML4-ALK) Are you able to consistently meet the recommended turnaround time goal of 10 working days for EGFR and EML4-ALK testing in non-small cell lung carcinomas? Yes 18 78.3 No 5 21.7 Abbreviations: ALK, anaplastic lymphoma receptor tyrosine kinase (anaplastic lymphoma kinase);BRAF, B-Raf proto-oncogene, serine-threonine kinase (v-raf murine sarcoma viral oncogene homolog B); EGFR, epidermal growth factor receptor;EML4, echinoderm microtubule-associated protein-like 4;KRAS, Kirsten rat sarcoma viral oncogene homolog;MMR, mismatch repair protein;MSI, microsatellite instability; ROS1, ROS protooncogene 1, receptor tyrosine kinase (c-ros oncogene 1). (a) Multiple responses allowed. Table 9. Laboratory Practices Related to Small Biopsy and Cytology Specimens No. Percentage Do you routinely use techniques to preserve tissue for molecular testing for small biopsy and FNA specimens? Yes, lung tumor biopsy specimens are 12 52.2 handled differently No, all specimens are handled in 11 47.8 basically the same manner If applicable, how are lung tumor biopsy specimens handled differently? (a) (n = 12) Additional unstained slides are 9 75.0 initially cut from biopsy specimens Fewer initial levels are made compared 3 25.0 to other biopsy specimens Immunohistochemical study is typically 2 16.7 limited to a certain number of stains Additional unstained FNA smears are made 1 8.3 Other 4 33.3 Other category submitted text response Dual stains used (1) EGFR ordered before immunohistochemistry (1) For lung biopsy specimens, what is the number of additional unstained slides initially cut? 3 1 11.1 5 4 44.4 6 2 22.2 9 1 11.1 15 1 11.1 For lung biopsy specimens, what is the typical limit of the number of slides for immunohistochemical study? 2 1 50.0 3 1 50.0 For fine-needle aspiration specimens, does your department provide preliminary adequacy assessment? Yes, routinely 21 91.3 No, we do not provide on-site assessment 2 8.7 Abbreviations: EGFR, epidermal growth factor receptor; FNA, fine-needle aspiration. (a) Multiple responses allowed. Abbreviations: ALK, anaplastic lymphoma receptor tyrosine kinase (anaplastic lymphoma kinase);BRAF, v-raf murine sarcoma viral oncogene homolog B; EGFR, epidermal growth factor receptor;EML4, echinoderm microtubule-associated protein-like 4; HER2-ERBB2, v-erb-b2 avian erythroblastic leukemia viral oncogene homolog 2 (human epidermal growth factor receptor 2);K-T, v-kit Hardy-Zuckerman 4 feline sarcoma viral oncogene homolog;KRAS, Kirsten rat sarcoma viral oncogene homolog;MMR, mismatch repair protein;MSI, microsatellite instability; RET, RET proto-oncogene. Table 10. Laboratory Practices Related to Molecular Testing Location In-House Send-Out Where are each of the No. Percentage No. Percentage following molecular tests performed in your practice? MMR immunohistochemistry 12 54.5 10 45.5 HER2/ERBB2 fluorescence 9 39.1 14 60.9 in situ hybridization KRAS mutation analysis 8 34.8 15 65.2 BRAF mutation analysis 7 30.4 16 69.6 EGFR mutation analysis 5 21.7 18 78.3 ALK (EML4-ALK) fluorescence 5 21.7 18 78.3 in situ hybridization MSI polymerase chain reaction 5 23.8 16 76.2 K/T mutation analysis 2 11.1 16 88.9 RET mutation analysis 0 0.0 17 100.0 Table 11. Laboratory Practices Related to Technical Aspects of Molecular Testing No. For in-house molecular testing on solid tumors, do you perform microdissection of specimens for molecular testing? (a) (n = 22) Yes, manual microdissection 6 Yes, laser capture microdissection 1 No, do not perform microdissection 5 of specimens N/A, we do not perform in-house molecular 10 testing on solid tumors If applicable, which personnel type typically performs laser capture microdissection of specimens? Personnel other than 1 histotechnologist/histotechnician, pathologist, or pathologist assistant If applicable, which personnel type typically performs manual microdissection of specimens? Histotechnologist/histotechnician 2 Personnel other than 5 histotechnologist/histotechnician, pathologist, or pathologist assistant For in-house molecular testing on solid tumors, does your molecular report include an assessment of the amount of tumor in the specimen? Yes 5 No 7 If yes, how is an assessment of the amount of tumor in the specimen reported? (a) (n = 5) Percentage of tumor cellularity is reported 4 Other 1 For in-house molecular testing on solid tumors, which of the following best reflects how your laboratory typically handles the 14-day rule for Medicare patients? Rule is disregarded; testing is initiated 6 at the time of request Request is held; testing is initiated 2 after appropriate time interval has passed Other 3 Other category submitted text response Immunohistochemistry performed in-house, no requisition (1) Rule is followed; testing initiated upon request (1) Test is performed as soon as possible (1) For send-out molecular testing on solid tumors, which of the following best reflects how your laboratory typically handles the 14-day rule for Medicare patients? Rule is disregarded; testing is initiated at 9 the time of request Request is held; material is sent out after 8 the appropriate time interval has passed N/A, we do not send out molecular tests 1 on solid tumors Reference laboratory holds the specimen 1 then runs the test at the appropriate time Other 3 Other category submitted text response Ask oncologist if result is needed (1) Rule is followed; testing initiated upon request (1) Percentage For in-house molecular testing on solid tumors, do you perform microdissection of specimens for molecular testing? (a) (n = 22) Yes, manual microdissection 27.3 Yes, laser capture microdissection 4.5 No, do not perform microdissection 22.7 of specimens N/A, we do not perform in-house molecular 45.5 testing on solid tumors If applicable, which personnel type typically performs laser capture microdissection of specimens? Personnel other than 100 histotechnologist/histotechnician, pathologist, or pathologist assistant If applicable, which personnel type typically performs manual microdissection of specimens? Histotechnologist/histotechnician 28.6 Personnel other than 71.4 histotechnologist/histotechnician, pathologist, or pathologist assistant For in-house molecular testing on solid tumors, does your molecular report include an assessment of the amount of tumor in the specimen? Yes 41.7 No 58.3 If yes, how is an assessment of the amount of tumor in the specimen reported? (a) (n = 5) Percentage of tumor cellularity is reported 80.0 Other 20.0 For in-house molecular testing on solid tumors, which of the following best reflects how your laboratory typically handles the 14-day rule for Medicare patients? Rule is disregarded; testing is initiated 54.5 at the time of request Request is held; testing is initiated 18.2 after appropriate time interval has passed Other 27.3 Other category submitted text response Immunohistochemistry performed in-house, no requisition (1) Rule is followed; testing initiated upon request (1) Test is performed as soon as possible (1) For send-out molecular testing on solid tumors, which of the following best reflects how your laboratory typically handles the 14-day rule for Medicare patients? Rule is disregarded; testing is initiated at 40.9 the time of request Request is held; material is sent out after 36.4 the appropriate time interval has passed N/A, we do not send out molecular tests 4.5 on solid tumors Reference laboratory holds the specimen 4.5 then runs the test at the appropriate time Other 13.6 Other category submitted text response Ask oncologist if result is needed (1) Rule is followed; testing initiated upon request (1) Abbreviation: N/A, not applicable. (a) Multiple responses allowed. Table 12. Laboratory Practices Related to Lynch Syndrome Screening No. Percentage For colorectal carcinomas, how does your practice typically handle requests for Lynch syndrome screening (MMR/MSI testing) by IHC and PCR? No algorithm, IHC and PCR are typically 6 26.1 performed concurrently Depends on clinician/geneticist request 6 26.1 Algorithm starting with IHC, with reflex PCR 5 21.7 Depends on pathologist 2 8.7 Algorithm starting with PCR, with reflex IHC 1 4.3 Other 3 13.0 Other category submitted text response IHC no reflex PCR (2) IHC performed first, then reflex PCR if applicable (1) How are patients typically selected for Lynch All (or nearly all) patients with colorectal 11 47.8 cancer are tested Depends on clinician request 7 30.4 Bethesda guidelines 1 4.3 Amsterdam criteria 0 0.0 Other 2 8.7 Other category submitted text response Amsterdam criteria, Bethesda guidelines, clinician request and findings (1) Histology with/without Bethesda guidelines (1) Do not know 2 8.7 For Lynch syndrome screening in colorectal carcinomas, does your department require informed consent from the patient for MMR/MSI testing by IHC and/or PCR? Yes 2 8.7 No 18 78.3 Do not know 3 13.0 Does your department perform Lynch syndrome screening tests on colorectal biopsy material? Yes 12 52.2 No 11 47.8 If yes, how is this typically handled? (a) (n = 12) IHC is performed 12 100.0 PCR is performed 4 33.3 If yes and PCR is performed, what is typically used for the patient's normal DNA? (n = 4) Normal tissue is microdissected from the 3 75.0 tumor biopsy block Additional normal tissue is purposely 1 25.0 obtained at the time of biopsy Abbreviations: IHC, immunohistochemistry;MMR, mismatch repair protein;MSI, microsatellite instability; PCR, polymerase chain reaction. (a) Multiple responses allowed.