Root Cause Analysis of Problems in the Frozen Section Diagnosis of In Situ, Minimally Invasive, and Invasive Adenocarcinoma of the Lung

Abstract

Context.-Frozen sections can help determine the extent of surgery by distinguishing in situ, minimally invasive, and invasive adenocarcinoma of the lung.Objective.-To evaluate our experience with the frozen section diagnosis of these lesions using root-cause analysis.Design. Results.-There were no false-positive diagnoses of malignancy. Frozen section errors and deferrals were identified in 12.1% (27 of 224) and 6.3% (14 of 224) of the cases, respectively. Significantly more errors occurred in the diagnosis of in situ and minimally invasive adenocarcinoma than in the diagnosis of invasive adenocarcinoma (P , .001). Frozen section errors and deferrals were twice as frequent in lesions smaller than 1.0 cm (P ¼ .09). Features significantly associated with errors and deferrals included intraoperative consultation by more than one pathologist (P ¼ .003) and more than one sample of frozen lung section (P ¼ .001). Inflammation with reactive atypia, fibrosis/scar, sampling problems, and suboptimal quality sections were identified in 51.2% (21 of 41), 36.6% (15 of 41), 26.8% (11 of 41), and 9.8% (4 of 41) of the errors and deferrals, respectively (more than one of these factors was identified in some cases). Frozen section errors and deferrals had significant clinical impact in only 4 patients (1.8%); each had to undergo completion video-assisted thoracoscopic lobectomy less than 90 days after the initial surgery.Conclusions.-The distinction of in situ from minimally invasive adenocarcinoma is difficult in both frozen and permanent sections. We identified several technical and interpretive features that likely contributed to frozen section errors and deferrals and suggest practice modifications that are likely to improve diagnostic accuracy.

Using frozen section to identify histological patterns in stage I lung adenocarcinoma of ≤3 cm: accuracy and interobserver agreement

“…With regard to AIS, MIA, and LPA, 79% of the interpretations using frozen sections were accurate for LPA; however, only 41.3% were accurate for MIA, and more than half (52%) of the interpretations were overdiagnoses of invasive adenocarcinoma. Our findings concurred with the recent study by Walts et al . In their study, only 46% of MIAs were accurately recognized in frozen sections, and there were 9% deferrals and 46% frozen section errors in cases of MIA.…”

Diagnosis of lung adenocarcinoma in situ and minimally invasive adenocarcinoma from intraoperative frozen sections: an analysis of 136 cases

“…Walts and Marchevsky10 reported that section quality and the diagnostic experience of the pathologist were significant factors that affected the accuracy of diagnoses based on frozen sections. For these reasons, some modified procedures were used in our study.…”

“…Walts and Marchevsky10 found that intraoperative frozen sections had a much lower diagnostic accuracy for AIS and MIA than invasive adenocarcinoma (50.68% vs 96.69%). In a retrospective analysis of pulmonary nodules with a diameter <1.5 cm from 183 patients, Marchevsky et al 11 found that the sensitivity of diagnoses based on intraoperative frozen sections was quite low for tumours with a diameter <1.1 cm.…”

Early lung cancer with lepidic pattern

“…However, in the strict sense, the definitive delineation of microinvasive foci is not critical from the clinical viewpoint (see below). intraoperative consultation of more than one pathologist necessary; more than one sample of frozen sections needed) as demonstrated by root cause analysis of 224 consecutive cases [34]. In addition, Yeh et al [33 & ] demonstrated that not only invasion per se but also different stromal invasion categories in combination with specific ADC patterns were associated with variations in patient survival.…”