3.1 Basic patient information

This study collected clinical and pathological information from 59 patients, including 37 males and 22 females, with an average age at surgery of 60.54 ± 10.67 years (Table 1). To investigate the expression of Ezrin in PDCs, paraffin blocks from these patients were respected and stained with H&E and IHC. Due to the limitations of previous wax block aging and fixation embedding techniques, the final IHC cohort included 29 males and 16 females (Table 2). Follow-up was performed based on overall survival (OS). Of the 59 patients, 7 were lost to follow-up and 6 patients died. The follow-up period ranged from 15 months to 12 years, with an average of 5.6 years.

Table 1 Demographic, clinical, and pathological characteristicsTable 2 The correlation between PDC-EMS and histopathological features

Statistical analysis revealed that more patients had right colon cancer and rectal cancer, with 24 cases (40.7%) of right colon cancer and 32 cases (54.2%) of rectal cancer. Only 3 cases (5.1%) were in left colon. From the perspective of gross morphology, there were 15 cases of protruding type, 43 cases of ulcerative type, and only 1 case of infiltrative type. According to the WHO grading standards, G1 and G3 tumors accounted for 16.9% and 11.9%, respectively, while G2 tumors accounted for 71.2 %. There were 6 LVI positive cases, accounting for 10.2%, and 13 PNI positive cases, accounting for 2.0 %.

According to TNMAJCC staging, there were 11 cases (18.6%) in stage I, 24 cases (40.7%) in stage II, and 24 cases (40.7%) in stage III. By pT staging, pT2, pT3, and pT4 accounted for 23.7%, 74.6%, and 1.7%, respectively. By pN staging, pN0, pN1, and pN2 were found in 59.3%, 37.3%, and 3.4% of the cases, respectively.

3.2 Correlation between PDC grading and other histopathological features

The PDC and TB grades were assessed using a three-tiered system (Table 3) and a two-tiered system (Table 4), respectively, and the relationships between the two grades and other clinicopathological features were summarized. The representative pathological characteristics of PDC and TB on H&E sections are shown in Fig. 1.

Table 3 The correlation between PDC and TB grading and histopathological features (three-tiered system)Table 4 The correlation between PDC and TB grading and histopathological features (two-tiered system)

This study found that there was a significant correlation between PDC grading and TB grading. Based on the number of PDCs, 12, 28, and 19 tumors were classified as PDC-G1, PDC-G2, and PDC-G3, respectively. The PDC grading based on the three-tiered system demostrated a significant positive correlation with TB grading (P = 0.003). Simultaneously, using the two-tiered system, PDC-G1 and PDC-G2 were grouped as low-grade PDC (a total of 30 tumors), while PDC-G3 was considered high-grade. Notably, the PDC grading based on the two-tiered system had an even more significant correlation with TB grading (P < 0.0001). These findings were in accordance with previous reports [15]. Meanwhile, the current study suggested that the two-tiered system for PDC and TB grading seemed simpler and more sensitive than the three-tiered system (Tables 3 and 4).

To validate the correlation between PDC and TB grading and some traditional pathological characteristics, a comparative analysis was conducted. The data indicated that PDC grading based on the three-tiered system had significant correlation with TNMAJCC stage, N stage, and gross morphology (P values were 0.003, 0.015, and 0.006, respectively), with a trend toward correlation with T stage (P = 0.08). Similarly, PDC grading based on the two-tiered system also showed significant correlation with TNMAJCC stage and N stage (P values were 0.0343 and 0.0192, respectively). Additionally, TB grading (both three-tiered and two-tiered) was significantly correlated with TNMAJCC stage and N stage (P values were 0.025, 0.011, 0.0048, and 0.0006, respectively), with the correlation between PDC grading and TNMAJCC stage being more significant. Furthermore, the data indicated that PDC grading (two-tiered) had a trend toward correlation with LVI and PNI (P values were 0.0566 and 0.0586, respectively), and that TB grading (two-tiered) also had a trend toward correlation with PNI (P = 0.0589). These results indicated that PDC grading (regardless of the grading method), was comparable to traditional histological indicators and the novel TB grading system.

Meanwhile, to explore the correlation between PDC and auxiliary indicators (TSR and E-cadherin), our group conducted relevant studies. As shown in Supplementary Tables 1 and Supplementary Table 2, with the increase of TB grade, the proportion of tumors with TSR < 1 and negative E-cadherin expression gradually increased, consistent with previous reports [27, 28]. Besides, when the stroma was denser (TSR < 1), E-cadherin expression showed a decreasing trend, exhibiting a significant statistical significance (P = 0.002) (Supplementary Table 1). TSR also showed some trend toward correlation with TNMAJCC stage, although there was no significant difference between the two parameters. As expected, E-cadherin expression decreased as TNMAJCC stage increased (P = 0.038), consistent with the results of previous study [11]. Furthermore, our results showed that the expression level of E-cadherin was negatively correlated with PNI (P = 0.001) and had significant statistical significance with pN stage (P = 0.021), but no statistical significance was found with LVI (Supplementary Table 2).

3.3 Correlation between PDC-EMS and other histopathological features

To assess the distribution trend of Ezrin-positive cell in PDC, IHC was applied to mark Ezrin-positive cells. Meanwhile, we developed a novel simplified method known as the PDC-EMS. The specific PDC-EMS scoring diagram is illustrated in Fig. 3.

According to our findings, PDC-EMS had a significant correlation with traditional pathological indicators. Of the 45 cases included in the TNMAJCC staging, there were 10 in stage I, 18 in stage II, and 17 cases in stage III, with mean PDC-EMS values of 1.64 ± 0.53, 1.80 ± 0.48, and 2.40 ± 0.58, respectively (P = 0.001). Furthermore, the mean PDC-EMS also had a significant correlation with the pN stage (P = 0.003). Although the mean PDC-EMS increased gradually with pT stage, there was no statistical significance. Additionally, the WHO grading system also showed good accordance with PDC-EMS (P = 0.029). Interestingly, our data also suggested that PDC-EMS correlated with gross morphology (P = 0.033), especially in ulcerative lesions, where the PDC-EMS were significantly higher than polypoid lesions. Meanwhile, a high PDC-EMS suggested an increased risk of LVI (P = 0.014), but this scoring system did not show any associations with PNI, gender, or tumor location (Table 2).

To validate the performance of PDC-EMS in prognostic evaluation, we compared it with PDC and TB. The results showed that based on TB grading (three-tiered system), there were 13, 15, and 17 cases in TB1, TB2, and TB3, respectively, with corresponding mean PDC-EMS values of 1.51 ± 0.48, 2.12 ± 0.52, and 2.52 ± 0.59, respectively (P = 0.001). Similarly, based on PDC grading (three-tiered system), there were 8, 22, and 15 cases of PDC-G1, PDC-G2, and PDC-G3, respectively, with corresponding mean EMS values of 1.40 ± 0.37, 2.04 ± 0.67, and 2.25 ± 0.40, respectively. The mean PDC-EMS showed a positive correlation with PDC grading in both three-tiered system (P = 0.004).and two-tiered system (P = 0.047). Meanwhile, this effective predictability of PDC-EMS was also synchronous with TB grading, indicating its reliability for clinical prognosis (Table 2).

Finally, our group also examined the relationship between mean PDC-EMS and auxiliary pathological indicators and found no significant correlation with TSR and E-cadherin. However, as the proportion of tumor stromal components increased (TSR < 1) and E-cadherin level decreased, the mean PDC-EMS value showed no significant increasing trend (Table 2).

3.4 The value of various pathological parameters on survival prognosis

To investigate the correlation between various pathological indicators and patient survival in the real-world setting, the research group employed univariate Kaplan-Meier analysis and multivariate Cox regression survival analysis. The univariate results indicated that the PDC-EMS grading demonstrated a superior prognostic value (P = 0.0081) for survival (Fig. 6), showing that the PDC-EMS-high group had a shorter survival time than the PDC-EMS-low group. Additionally, pT staging (P = 0.0056) confirmed its clinical prognostic value, whereas other traditional pathological indicators such as pN staging, TNM stage, WHO grading, LVI, and PNI, did not show a significant correlation with clinical prognosis. Notably, PDC grading (three-tiered system) was identified as a risk factor for prognosis (P = 0.018). At the same time, both PDC grading (two-tiered system) and TB grading (two-tiered system) showed a tendency to become risk factors for predicting postoperative survival (P values were 0.059 and 0.055, respectively) (Fig. 7). These results indicated that PDC and TB grading offer greater advantages over traditional pathological indicators regarding clinical prognosis, and that PDC-EMS system proves to be superior to both novel and traditional pathological indicators.

Fig. 6

Overall survival (OS) of stages I–III CRCs by PDC-EMS (high or low)

Furthermore, multivariate Cox regression analysis was conducted to identify risk factors affecting postoperative survival in CRC. When eight factors were included in the regression analysis, significant statistical significance was found (P = 0.029), with E-cadherin and PNI as independent influencing factors (P values were 0.017 and 0.015, respectively) (Table 5). When only PDC grading, pT staging, and LVI were included, it was found that PDC grading was an independent factor affecting postoperative survival (P = 0.026) (Table 6).

Table 5 Pathological factors and colorectal tumor patient prognosis: cox multivariate regression analysisTable 6 Multivariate Cox regression analysis of factors affecting postoperative survival in CRC