ABSTRACT AJCP /ORIGINAL ARTICLE

Transcription

1 Assessment of the Intraoperative Consultation Service Rendered by General Pathologists in a Scenario Where a Well-Defined Decision Algorithm Is Followed Mahmoud A. Khalifa, MD, PhD, 1 Sherine Salama, MD, 3 Rachel I. Vogel, PhD, 2 Molly E. Klein, MD, 1 James Richter, MD, 1 Tanya Pulver, MD, 2 Sally A. Mullany, MD, 2 and Boris Winterhoff, MD 2 From the 1 Department of Laboratory Medicine and Pathology and 2 Department of Obstetrics, Gynecology and Women s Health, Division of Gynecologic Oncology, University of Minnesota, Minneapolis; and 3 Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Canada. Key Words: Intraoperative consultation; Frozen sections; Anatomic pathology quality; Decision algorithms; General pathologists; Subspecialization in pathology Am J Clin Pathol March 2017;147: DOI: /AJCP/AQW223 ABSTRACT Objectives: Intraoperative consultation (IOC) remains an area of general practice even within subspecialized pathology departments. This study assesses the IOCs rendered in a general pathology setting where surgeons integrate these results in a well-defined algorithm, developed with the input of specialized pathologists. Methods: The surgical decisions to perform lymphadenectomy in patients with endometrial adenocarcinoma operated on at our institution between January 2003 and June 2015 as a result of the IOC assessment of tumor size, histologic grade, and depth of invasion in the hysterectomy specimen were analyzed. Results: Frozen section (FS) was examined in 801 cases. In comparison to permanent section analysis, FS International Federation of Gynecology and Obstetrics (FIGO) grade had an overall accuracy of 0.95 (95% confidence interval [CI], ). The FS depth of invasion had an overall accuracy of 0.92 (95% CI, ). FIGO grade was not documented in 47.8%, the depth of myometrial invasion in 45.2%, and tumor size in 41.8% of the pathology reports. Conclusions: The high omission rate of the needed parameters by the general pathologists would question their overall understanding of the paradigm shift intended by this algorithm. Possible explanations of this phenomenon and potential solutions are discussed. Subspecialization has existed in pathology for a long time as a natural response to the increasing subspecialization in surgical and oncologic specialties. In addition, subspecialization offered a solution to the dilemma of increased workload, whether due to increased volumes or increased reporting requirements for various cancer and noncancer cases. 1 Currently, most North American academic departments of pathology have adopted some degree of subspecialization in their practice, with an increasing move to broaden its adoption into the community setting. 2,3 With the wealth of experience accumulated by subspecialized academic pathologists, management decision algorithms have been increasingly dependent on the more detailed level of expertise rendered by this model. As a result, many of the more sophisticated algorithms currently in practice have interpretive data elements that have been designed, validated, and implemented by subspecialized pathologists who are experts in their specific field. There are rather limited published data on the feasibility, accuracy, or clinical outcome when these interpretive decision points are executed by general pathologists in scenarios where well-defined multidisciplinary decision algorithms are followed. For example, irrespective of where on the subspecialization spectrum a pathology practice stands, intraoperative consultation (IOC) remains an area of general practice even within some of the most subspecialized departments. The University of Minnesota has a large academic group of gynecologic oncologists who have incorporated, into their 322 Am J Clin Pathol 2017;147: American Society for Clinical Pathology, All rights reserved. For permissions, please journals.permissions@oup.com

2 surgical treatment of type I endometrial cancer, the endometrial cancer staging algorithm developed jointly by the surgeons and subspecialized pathologists at Mayo Clinic. 4-6 According to this algorithm, patients who have International Federation of Gynecology and Obstetrics (FIGO) grade 1 or 2 endometrioid endometrial adenocarcinoma with myometrial invasion 50% or less, greatest surface tumor dimension 2 cm or less, and no intraoperative evidence of macroscopic disease outside the uterus can be surgically treated with hysterectomy alone. Using this algorithm, the decision to omit lymphadenectomy in this select group of patients with low-risk disease is made based on the preoperative determination of the tumor FIGO grade with subsequent confirmation based on frozen section (FS) analysis by the pathologist intraoperatively of FIGO grade, tumor size, and depth of myometrial invasion. The clinical applicability and appropriateness of this approach to patient care have been further validated by others. 7,8 IOC service at the University of Minnesota has traditionally been provided by general pathologists. This situation has created an excellent opportunity to study the appropriateness of rendering IOC by general pathologists when interpretive decision points are integral to a well-defined management algorithm. The purpose of this audit was to evaluate outcomes after the adoption of a well-defined multidisciplinary decision algorithm, designed by subspecialized pathologists and executed by a group of general pathologists. Materials and Methods CoPath database (Sunquest, Tucson, AZ) was implemented in the Department of Pathology at the University of Minnesota in late 2002 and was used to identify cases for this practice audit analyzing all endometrial cancer cases between January 2003 and June We searched the tissue part type for hysterectomy and filtered the natural language search for adenocarcinoma. Then we further filtered our search for hysterectomy specimens with endometrial adenocarcinoma where an IOC was rendered. Nonuterine malignancies with metastasis to the endometrium or nonendometrioid endometrial cancers including carcinosarcomas were excluded. We identified 897 hysterectomy specimens that met the inclusion criteria. The IOC focused on nonendometrial sites in 82 cases (excluded) and was rendered on the endometrial tumor in the remaining 815. In 14 of these 815 cases, the IOC was rendered based on gross examination only. The remaining 801 hysterectomy cases had FS and were included in this audit. Data abstraction occurred from CoPath pathology reports. FIGO tumor grade and depth of myometrial invasion reported intraoperatively were compared with those in the permanent sections. Tumor size documented intraoperatively was also captured but was not included in a comparison. When tumor size was captured as tumor replaces the entire endometrium or as >2 cm, the case was counted as an actual reporting of tumor size. Other data points collected included the surgeon, the IOC pathologist, the final pathologist, pelvic lymph nodes (total number and number of positive), periaortic lymph nodes (total number and number of positive), and the FIGO tumor grade in the preoperative specimen diagnosis when available. Since the algorithm studied handled FIGO grades 1 and 2 collectively in the same group, our accuracy calculations lumped these two grades together. Surgical and patient outcomes were obtained for patients with discrepant IOC and permanent results from the electronic medical records. Point estimates of the proportions and Clopper-Pearson 95% confidence intervals (CIs) are presented for each scenario of the FIGO tumor grade and depth of myometrial invasion comparing the sensitivity, specificity, false negatives, and accuracy of the IOCs vs permanent section. Data analyses were conducted using SAS 9.3 software (SAS Institute, Cary, NC). Results This audit included 801 hysterectomy specimens where IOC was rendered on the endometrial tumor by means of FSs; 14 cases had gross examination only. Surgeries were performed by 18 gynecologic oncologists, and the IOCs were rendered by 16 general pathologists. Of the 801 cases with FS, the pathologists did not document the tumor grade in the IOC portion of their final report in 383 (47.8%) cases. In 64 cases, the pathologist either did not diagnose adenocarcinoma on the FS submitted (58 cases) or diagnosed an endometrioid adenocarcinoma but deferred its FIGO grade (six cases) to permanent section. All 64 cases were FIGO grade 1 on permanent section. Findings regarding the documented FS FIGO tumor grade in the remaining 354 cases are summarized in Table 1. The FS tumor FIGO grade had an overall accuracy of 0.95 (95% CI, ). FIGO tumor grade 3 was undercalled in 11 patients. Of these, the surgeon proceeded with the staging lymphadenectomy in one patient due to a bulky tumor (3.0 cm) in the setting of a preoperative diagnosis of highgrade histology. The 10 remaining patients did not receive staging lymphadenectomy. Two of these 10 patients died of disease, 14 and 23 months postoperatively. A third patient had a lymph node recurrence 27 months postoperatively. The remaining seven patients had no evidence of disease after a median follow-up of 74 months. Of the 801 cases with FS, the pathologists did not document the depth of myometrial invasion in the IOC portion of American Society for Clinical Pathology Am J Clin Pathol 2017;147: Downloaded 323 from

3 Khalifa et al /RENDERING IOC BY GENERAL PATHOLOGISTS Table 1 Summary of Correlation Between Tumor Grade Reported Intraoperatively vs That on Sections IOC Report Grade 1 or 2 Grade 3 Total FIGO grade 1 or FIGO grade Total FIGO, International Federation of Gynecology and Obstetrics; IOC, intraoperative consultation. Table 3 Summary of Correlation Between Depth of Invasion Reported Intraoperatively by Gross Examination Only vs That on Sections IOC Report (Gross) <50% >50% Total <50% >50% Total IOC, intraoperative consultation. Table 2 Summary of Correlation Between Depth of Invasion Reported Intraoperatively by FS vs That on Sections IOC Report (FS) <50% >50% Total <50% >50% Total patients had tumor size less than 2 cm in the greatest dimension, measured intraoperatively. Of these, one patient had more than 50% myometrial invasion, and another had a FIGO grade 3 tumor. The remaining 41 patients with tumors less than 2 cm, myometrial invasion less than 50%, and FIGO grades 1 or 2 all had staging lymphadenectomy. FS, frozen section; IOC, intraoperative consultation. their final report in 362 (45.2%) cases. In 58 cases, the pathologist did not diagnose adenocarcinoma on the FS submitted. Findings regarding the documented FS of the depth of invasion in the remaining 381 cases are summarized in Table 2. Of the 14 cases where the IOC was rendered by gross examination only, the pathologists did not document the depth of invasion in five cases. Findings regarding the documented depth of invasion by gross examination only are summarized in Table 3. The FS depth of myometrial invasion had an overall accuracy of 0.92 (95% CI, ). Deep myometrial invasion (>50%) was undercalled in 25 cases. Staging lymphadenectomy was performed in 20 of these patients due to large tumor size or a FIGO grade 3 diagnosis. The remaining five patients had both preoperative and FS diagnoses reflecting FIGO grade 1 tumors; therefore, lymphadenectomy was not performed despite intraoperative documentation of tumor sizes greater than 2 cm. None of these five patients had tumor recurrence after a median follow-up of 70 months. Of the total 815 cases with IOC (gross or FS), the pathologists did not document the tumor size in 341 (41.8%) cases. The tumor size documented in the final report was measured intraoperatively. Accordingly, tumor size was not further analyzed, but the rate of its omission was recorded. Staging lymphadenectomy was performed on 553 (69.0%) of the 801 patients with FS diagnoses. Positive pelvic lymph node metastases were detected in 44 patients and periaortic lymph nodes were positive in 18 patients. Fifteen of the 18 patients with positive periaortic lymph nodes also had positive pelvic lymph nodes, while in the remaining three patients, pelvic lymph nodes were negative. Forty-three Discussion For decades, the depth of myometrial invasion by endometrial adenocarcinoma has been recognized as a predictor for lymph node metastases. 9 Consequently, it has become standard practice to obtain pathologic intraoperative assessment of endometrial adenocarcinoma hysterectomy specimens, by FS and gross examination, 10 to guide surgical decision making by the gynecologic oncologist regarding indication for staging lymphadenectomy. Investigations led by Mariani and colleagues 4,5 at the Mayo Clinic intended to identify low-risk endometrial cancers in which lymphadenectomy was not necessary. 6 This research, which has been further validated by others, focused on streamlining the surgical approach and optimizing adjuvant therapy to improve patient outcomes. 7,8 It was developed with the input of highly specialized pathologists who worked closely with their clinical colleagues to systematically evaluate hysterectomy specimens. Although there is an acknowledged estimated intrinsic failure rate of 2% in the FIGO grading of endometrial cancer, this algorithm became widely adopted across many medical centers in North America. 11 For pathologists, this work has shifted their paradigm from a single metric capture, which was rather objective (ie, depth of myometrial invasion), to the provision of multiple interpretive data elements that have become integral to a welldefined management algorithm. These elements included the preoperative tumor grade, which may or may not be reviewed in the center where surgery is conducted; the FS confirmation of the tumor grade; the intraoperative depth of myometrial invasion; and the measurement of tumor size. In this audit, the 16 general pathologists who carried out this 324 Am J Clin Pathol 2017;147: American Society for Clinical Pathology 324

4 task from January 2002 to June 2015 did not document the tumor grade in 47.8%, the depth of myometrial invasion in 45.2%, and the tumor size in 41.8% of cases; all these are metrics needed to execute the well-defined multidisciplinary decision algorithm. The diagnostic accuracy of the pathologists FSs was rather appropriate. While the general pathologists were able to deliver the specific interpretive tumor metrics with high precision, the high omission rate of the needed parameters would question their overall understanding of the paradigm shift introduced by this algorithm. Several quality assurance publications have addressed the reliability and/or accuracy of FS assessment of FIGO tumor grade and depth of myometrial invasion compared with the findings reported on permanent sections with slightly conflicting results Our data showed accuracy of FS reporting during the period audited, falling within the published range and exceeding that reported in several of these series. We conclude that the general pathologists diagnostic skills pertaining to the specific metrics that were deemed necessary by the subspecialized team were supported by the accuracy rates of their results in this audit. However, their consistency in documenting these findings is not supported by the high omission rates witnessed in their reports. There are several possible explanations for this phenomenon. First, there may be a generalized lack of awareness among general pathologists regarding the paradigm shift from the need to report a single objective metric to the provision of multiple interpretive data elements integral to a welldefined decision algorithm. Findings from this audit indicate a need to develop reproducible, efficient tools to help facilitate the adoption of similar surgicopathologic algorithms among practicing pathologists. Checklists could be an easy solution in which a general pathologist is reminded, in this scenario for instance, to document the FIGO tumor grade, depth of myometrial invasion, and tumor size consistently on every endometrial cancer hysterectomy specimen intraoperatively. There are ongoing clinical situations where subspecialized pathologists, working closely with their clinical specialists, have developed and will continue to develop multidisciplinary interpretation-dependent management algorithms. Designing companion checklists to be used with the wide adoption of these models by general pathologists could be an effective strategy. Second, another factor that would contribute to the high omission rate seen in this particular audit is the fact that, while the Mayo Clinic s approach was a well-defined multidisciplinary decision algorithm, it was not consistently adopted over time by the 18 members of the gynecologic oncology team, and it has been implemented only as an overarching divisional policy in the past 2 years. Consequently, the various data elements needed in order for this model to work were not consistently requested from pathology. Alternatively, these pathologic criteria may have been communicated to the surgeon verbally but not necessarily documented in the actual pathology report. In the setting of a complex disease process in which surgical decision making reflects a multidisciplinary decision algorithm, such inconsistencies in practice need to be avoided. As a patient safety initiative, we have implemented a workflow modification into our IOC reporting whereby the pathologist directly enters his or her interpretation in CoPath and signs it out as an addendum that crosses from the laboratory information system to the electronic medical record prior to finalizing the pathology report. This project was led by one of the authors (M.E.K.) to improve communication with the surgeons in real time intraoperatively by adding the visual documentation of diagnoses to reduce the probability of miscommunication created by the verbal diagnosis alone. With more sophisticated programing, an organ-specific IOC checklist can be automatically inserted in the IOC section, filtered by the part type and keyword diagnosis (eg, adenocarcinoma), to remind the pathologist of all the necessary data elements needed in a given clinical scenario. The system can also block the sign-out and release of an incomplete checklist intraoperatively to ensure consistent reporting. In summary, this audit suggests that while the general pathologists were able to deliver the specific interpretive tumor metrics with high precision, a high omission rate of the needed parameters did exist. To improve the consistent reporting of these clinically relevant parameters, implementation of systemwide electronic checklists would ensure the standardized capture of these important pathologic metrics in an effort to improve patient outcomes and individualize treatments. Corresponding author: Mahmoud A. Khalifa, MD, PhD, Dept of Laboratory Medicine and Pathology, University of Minnesota, 420 Delaware St, SE MMC 76, Minneapolis, MN 55455; mkhalifa@umn.edu. Disclaimer: This work was presented in part at the USCAP Annual Meeting, March 2016, Seattle, Washington. References 1. Black-Schaffer WS, Young RH, Harris NL. Subspecialization of surgical pathology at the Massachusetts General Hospital. Am J Clin Pathol. 1996;106:S33-S Sarewitz SJ. Subspecialization in community pathology practice. Arch Pathol Lab Med. 2014;138: Groppi DE, Alexis CE, Sugrue CF, et al. Consolidation of the North Shore-LIJ Health System anatomic pathology services: the challenge of subspecialization, operations, quality management, staffing, and education. Am J Clin Pathol. 2013;140: American Society for Clinical Pathology Am J Clin Pathol 2017;147: Downloaded 325 from

5 Khalifa et al /RENDERING IOC BY GENERAL PATHOLOGISTS 4. Mariani A, Webb MJ, Keeney GL, et al. Low-risk corpus cancer: is lymphadenectomy or radiotherapy necessary? Am J Obstet Gynecol. 2000;182: Mariani A, Dowdy SC, Cliby WA, et al. Prospective assessment of lymphatic dissemination in endometrial cancer: a paradigm shift in surgical staging. Gynecol Oncol. 2008;109: Bogani G, Dowdy SC, Cliby WA, et al. Role of pelvic and para-aortic lymphadenectomy in endometrial cancer: current evidence. J Obstet Gynecol Res. 2014;40: Convery PA, Cantrell LA, Di Santo N, et al. Retrospective review of an intraoperative algorithm to predict lymph node metastasis in low-grade endometrial adenocarcinoma. Gynecol Oncol. 2011;123: Milam MR, Java J, Walker JL, et al. Nodal metastasis risk in endometrioid endometrial cancer. Obstet Gynecol. 2012;119: Creasman WT, Morrow CP, Bundy BN, et al. Surgical pathologic spread patterns of endometrial cancer: a Gynecologic Oncology Group study. Cancer. 1987;60S: Doering DL, Barnhill DR, Weiser EB, et al. Intraoperative evaluation of depth of myometrial invasion in stage I endometrial adenocarcinoma. Obstet Gynecol. 1989;74: Helpman L, Kupets R, Covens A, et al. Assessment of endometrial sampling as a predictor of final surgical pathology in endometrial cancer. Br J Cancer. 2014;110: Atad J, Weill S, Ben-David Y, et al. Intraoperative frozen section examination of myometrial invasion depth in patients with endometrial carcinoma. Int J Gynecol Cancer. 1994;4: Case AS, Rocconi RP, Straughn JM Jr, et al. A prospective blinded evaluation of the accuracy of frozen section for the surgical management of endometrial cancer. Obstet Gynecol. 2006;108: Sanjuan A, Cobo T, Pahisa J, et al. Preoperative and intraoperative assessment of myometrial invasion and histologic grade in endometrial cancer: role of magnetic resonance imaging and frozen section. Int J Gynecol Cancer. 2006;16: Kumar S, Medeiros F, Dowdy SC, et al. A prospective assessment of the reliability of frozen section to direct intraoperative decision making in endometrial cancer. Gynecol Oncol. 2012;127: Am J Clin Pathol 2017;147: American Society for Clinical Pathology 326