Atherosclerotic plaque samples obtained during carotid endarterectomy from 24 symptomatic patients were collected. The endarterectomy specimens were cut into parallel, transverse segments of 5‐mm thickness. Each alternating segment was snap‐frozen in liquid nitrogen and stored at −80°C, their flanking segments fixed for 24 h in formalin, decalcified for 4 h before processing and embedding in paraffin for histological evaluation.
Hematoxylin–eosin (H&E) stained plaque tissue was macroscopically preclassified for plaque stage before omics experiments. Segments were stratified into non‐IPH and IPH groups according to the absence or presence of IPH. For each CEA specimen and patient, two samples (one non‐IPH and one IPH) were collected for omics experiments (n = 24 at the start for each). However, one IPH and four non‐IPH samples did not pass the QC test and were excluded from analysis (see the following sections). Simultaneously with the microarray performance for the omics experiments, but prior to bioinformatic analysis, tissue was sectioned further for additional staining and computer‐aided quantitative measurement of plaque IPH. Based on this indepth reinspection, the three experienced pathologists (MJAPD, JCS, MJJG) agreed to remove one ambiguous “non‐IPH” sample, showing small but surface‐detached luminal fibrin clot, as this could represent either a surgery artefact or bona fide IPH. In three allegedly “non‐IPH” samples, quantitative morphometry detected minor IPH (0.43%, 0.40%, 0.33%, respectively, data not shown), which was overlooked in the preclassification. These three were recategorized as IPH after inspection by the pathologists. Due to the recategorization of samples from the non‐IPH to the IPH group, three pairs of the total 26 samples of the IPH group were from three patients, respectively. To explore the potential confounder effects of this adjustment, we performed hierarchical clustering of the samples based on the plaque traits (measured as described in Section 2.2). This analysis convincingly showed that there is no patient‐specific heterogeneity among the samples, as well as that plaque phenotype is dominant over sample origin (data not shown). The final cohort for this study therefore included transcriptomics, proteomics, and peptidomics for 16 non‐IPH and 26 IPH plaques (see Figure1A ,B for flow scheme of cohort build‐up). For detailed information on the patient cohort definition, see Table S1 .
Hematoxylin–eosin (H&E) stained plaque tissue was macroscopically preclassified for plaque stage before omics experiments. Segments were stratified into non‐IPH and IPH groups according to the absence or presence of IPH. For each CEA specimen and patient, two samples (one non‐IPH and one IPH) were collected for omics experiments (n = 24 at the start for each). However, one IPH and four non‐IPH samples did not pass the QC test and were excluded from analysis (see the following sections). Simultaneously with the microarray performance for the omics experiments, but prior to bioinformatic analysis, tissue was sectioned further for additional staining and computer‐aided quantitative measurement of plaque IPH. Based on this indepth reinspection, the three experienced pathologists (MJAPD, JCS, MJJG) agreed to remove one ambiguous “non‐IPH” sample, showing small but surface‐detached luminal fibrin clot, as this could represent either a surgery artefact or bona fide IPH. In three allegedly “non‐IPH” samples, quantitative morphometry detected minor IPH (0.43%, 0.40%, 0.33%, respectively, data not shown), which was overlooked in the preclassification. These three were recategorized as IPH after inspection by the pathologists. Due to the recategorization of samples from the non‐IPH to the IPH group, three pairs of the total 26 samples of the IPH group were from three patients, respectively. To explore the potential confounder effects of this adjustment, we performed hierarchical clustering of the samples based on the plaque traits (measured as described in Section 2.2). This analysis convincingly showed that there is no patient‐specific heterogeneity among the samples, as well as that plaque phenotype is dominant over sample origin (data not shown). The final cohort for this study therefore included transcriptomics, proteomics, and peptidomics for 16 non‐IPH and 26 IPH plaques (see Figure
Free full text: Click here
