92 Chapter 6 lling air spaces; and (iii) discohesive single cells in alveolar spaces117. The latter 2 criteria are not different from our ndings in this prospective study, and interestingly, the frequency reported for STAS is similar to the frequency of loose tissue fragments reported in this study. The concept of STAS as a speci c pattern of invasion was added to the 2015 WHO Classi cation of Lung Tumors112 and published with only one supporting reference (listed as “Forthcoming” in the WHO bibliography). Three-dimensional reconstruction studies demonstrating tumor islands connected to one another and with the index tumor at multiple points provides additional support. These observations suggest that tumor islands represent tumor extension through air spaces132 133. While some of our 7% loose fragments in block 1 may be explained by this 3-dimensional growth, the presence of benign loose fragments, that is, normal respiratory epithelium detached from its origin in these tissue blocks, argues against this possibility and instead supports the hypothesis that this nding is artifactual. Furthermore, STAS has recently been reported in SqCC133. Three-dimensional growth within alveolar spaces is a difficult concept to imagine given the nature of this tumor cell type. “STAS” reportedly has prognostic signi cance and correlates with aggressive tumor behavior and poor prognosis115 118 114 116 134, especially in sublobar resections117. However, one could argue that STAS-like artifact is simply seen in more discohesive tumors, that is more poorly differentiated tumors with fewer intercellular adhesions, such as micropapillary AdC, and this quality is responsible for the worse prognosis observed in these tumors. The presence of STAS in SqCC is also associated with poorly differentiated features such as necrosis, nuclear diameter and Ki67 labeling index133. If so, then STAS is simply an epiphenomenon with a biologic explanation rather than true tumor invasion. In addition, in our study there was no apparent increase of benign loose fragments in blocks 2 to 4, in line with the more cohesive nature of benign epithelial cells. Thus, it may be premature to recognize STAS as a morphologic pattern of lung cancer invasion. The concept of tumor invasion and lung carcinoma staging is already very complex, given the recent decision to reclassify lepidic-pattern AdC as AdC in situ, and add a super cially studied minimally invasive AdC to the morphologic classi cation. Our data suggests that STAS likely, in many instances, represents mechanical artifacts including knife spread during specimen prosection (STAKS). Our study has several limitations. Although it had a detailed protocol for tissue processing, the size of the tissue sample and proportion of tumor was not constant in the different formalin- xed paraffin-embedded blocks. This may have led to differences in amounts of normal lung tissue between the samples and therefore theoretically to a different number of detected artifactual lesions. Second, this study has no followup data, which precludes statistical analysis with regard to the possible prognostic relevance of the artifact vis-à-vis STAS. Lastly, this study was limited in that the AdC were mainly non-mucinous tumors. We offer no data or opinion on mucinous lung AdC and scattered detached tumor fragments.
RkJQdWJsaXNoZXIy MTk4NDMw