193 Pattern recognition in pulmonary AdC; a modified classification cells, the only potential way to handle the sideways pressure during iatrogenic collapse is the displacement of cellular material towards the alveolar lumen on the apical side. If this hypothesis is correct, it is possible that tumour cells may exhibit an increased apicobasal height in instances of more pronounced collapse. In the test set, 20 cases were suitable for measurements, while in the validation cohort 16 cases could be measured. The measurements of apicobasal tumour cell length and relative air surface are shown in Table 2. Table 2. The mean ± standard deviation (SD) of minimum tumour cell length, relative air surface and stroma is shown for Milan (test cohort) and Amsterdam cases (validation cohort). Cohort Minimum TC length Relative air surface Stroma Mean ± SD Mean ± SD Mean ± SD Test, n=20 8.3 ± 2.6 18.8 ± 14.6 18 ± 6.9 Validation, n=16 9.6 ± 2.9 23.2 ± 15.8 13.5 ± 6.2 The correlation coefficient between the relative amount of air and minimal tumour cell height in the Milan cohort’s test set showed an inverse relationship. Put differently, the smaller the amount of air (i.e., more pronounced collapse), the larger the apicobasal tumour cell length (r = -0.44; p < 0.001). In the independent OLVG cohort, this finding was validated (r = -0.42; p = 0.001). For the maximal tumour cell length, the correlation values were similar for both test and validation set, r = -0.43 (p < 0.001) and r = -0.39 (p = 0.002), respectively. The correlation coefficient for stromal area with minimal tumour cell length was not significant in both data sets (p values in test and validation set: 0.50 and 0.64, respectively). Reclassification of cases The clinical characteristics of the two cohorts that were tested are shown in table 3. Out of the 70 cases, 56 contained a scar (80%), of which 38 (54%) had a combination of biological collapse and fibrosis. The remaining cases showed biological collapse (n=11, 16%) or fibrosis only (n=7, 10.0%). Of the 10 AIS cases, 5 did not show scar signs and the remaining showed biological collapse (n=3) or a combination of biological collapse and fibrosis (n=2). In 48 of the 56 cases (86%) with a scar, tumour infiltration was seen within the scar. In all 7 “fibrosis only” scars invasion was present (100%), while this was the case in 35 of the 38 cases with a combined scar (92%). In 6 of the 11 (55%) “biologic collapsed only” scars, invasion was discerned. 14
RkJQdWJsaXNoZXIy MTk4NDMw