Clinicopathological characteristics, molecular alterations and prognosis of mucinous histology in colorectal adenocarcinoma patients

Mucinous adenocarcinoma (MAC) is conventionally diagnosed by WHO denition when the extracellular mucin is more than 50% of the tumour area. The study is aimed at analyzing the clinicopathological characteristics, mutation spectrum, and prognosis of the colorectal adenocarcinoma (CRC) with any mucinous proportion or feature. Clinical and pathological information for 50 patients were reviewed and recorded. Mutation analyses for exon 2–4 of KRAS gene and exon 15 of BRAF gene were performed by Sanger sequencing. Expression of DNA mismatch repairs (MMRs) and P53 proteins was evaluated by immunohistochemistry (IHC). Density of tumour inltrating lymphocyte (TIL) status was scored. We also evaluated the percentage of glands producing mucin and the morphology of the different tumor cells types in mucin pools. We retrospectively analyzed prognosis of 43 patients with Stage II-III MAC. The primary outcome was progression-free survival (PFS).


Results
The overall frequencies of KRAS and BRAF mutations were 37.9% and 4.4%, respectively. Patients with MAC exhibiting high levels of mucin were related to the increase of tumor diameter (P = 0.038), but were not associated with any of the other clinicopathological parameters. The proportion or variable morphology of mucinous component did not stratify PFS in Stage II-III tumours. It is interesting to note that male patients had lower TIL compared with female patients (P = 0.018). TIL-low tumors were also correlated with advanced stage (P = 0.041). TIL status was a strong independent predictor of PFS in stage II-III mucinous component tumours (p < 0.001). All four IV stage patients were also classi ed into TIL-low case. No parameters were independently associated with outcome after adjusting for tumour stage in multivariate analysis.
Conclusions TIL status could more accurately determine the biology of the MAC feature for appropriate management, irrespective of quantity or morphology of mucinous component.

Background
Mucinous adenocarcinoma (MAC) is considered a special histologic type of colorectal adenocarcinoma(CRC)with a substantial amount of extracellular mucin. The de ning features of this tumour have been based variously on the presence of ≥ 50% mucinous component of the tumour [1], the mucin pool may contain layers, acini, or cribriform sheets of malignant glands, or scattered individual signet ring cells. This cutoff, while traditional, is largely arbitrary and has not been rigorously evaluated.
MAC is a distinct subtype of CRC. Despite several studies comparing underlying clinicopathological differences between MAC and conventional CRC. Still, controversial results relating prognostic value of mucinous histology have been reported [2][3][4][5][6][7] [5]. A study by Inamura et al showed that even when less than 50% component of signet ring cells were present in the tumor, they could still serve as a poor prognostic indicator in CRC, independent of other clinicopathological features [9]. Song et al reported that MAC with more than 50% signet ring cell component and MAC with a greater amount of extracellular mucin were associated with poor progression-free survival (PFS) [10]. Since relatively few studies have analysed the clinicopathologic characteristics of MAC (including histological subtypes and percentage of mucinous component) in relation to patient outcome, we classi ed the MAC into the following: 1) three categories based on the proportion of extracellular mucin pools. 2) Several histological subtypes based on the mucinous components of different morphology.
In the present study, we explored prognostic interactions between clinico-molecular variables and tumour mucinous component and investigated the prognostic value of TIL status.

Patients
The medical records of 50 patients who underwent surgical resection at Xuhui District Center Hospital of Shanghai from 2017 to 2020 with a diagnosis of primary MAC. Archived formalin-xed para nembedded histopathology slides of these patients were retrieved for con rmation of diagnosis and molecular, immunohistochemistry (IHC) and histopathological studies. Patients who received any neoadju vant chemotherapy or radiotherapy were excluded. Written informed consent was acquired from all the enrolled patients. This study was approved by our institutional review board. Clinicopathological data were collected for analyses of the association with MAC subtypes sex, age, pathological cancer stage, histological type, tumor size, tumor location, lymph nodes involvement and tumor differentiation were included. Two experienced pathologists reviewed all MAC cases using a multi-viewing microscope and were in agreement regarding histological grading and subtyping.

Histopathologic evaluations
Depending on the proportion of extracellular mucin, tumors were separated into three categories: 1) low mucin, extracellular mucin occupying < 50 % of the tumour; 2) moderate mucin, 50-80 % of the tumour; 3) high mucin, ≥ 80 % of the tumour. In terms of the microscopic morphology of tumour cells in mucin pools, three types were de ned: 1) strip predominant type, characterised by attened single-cell layers or strips of tumour cells located along the edge of mucin pools; 2) cluster predominant type, denoted by oating tumour cells forming clusters, acini, or cribriform sheets; 3) predominant signet ring cell type, in which signet ring cells compose ≥ 50 % of tumour cells [8]. The number of poorly differentiated clusters (PDCs) in a single eld of highest activity was graded as G1 (< 5 clusters), G2 (5 to 9 clusters), or G3 (≥ 10 clusters) under an objective lens with a magni cation of ×20 according to previous article [11].
Mutation analysis DNA from para n-embedded tissue was extracted, and KRAS (exons 2-4), and BRAF (exon15) were ampli ed by polymerase chain reaction (PCR) using complementary DNA. Ampli ed products were assessed using the Sanger direct-sequencing method in forward and reverse directions to detect mutations.
Immunohistochemistry staining analysis Antibody were purchased from Talent Biomedical Technology Co.,Ltd. (Xiamen China). IHC staining for P53 was performed in routinely para n embedded specimens. Appropriate positive and negative controls were included in each run of IHC. Microsatellite instability was assessed using IHC staining for mismatch repair (MMR) proteins, including MLH1, PMS2, MSH2 and MSH6. Loss of expression of any of the MMR proteins was considered when there was no corresponding IHC staining in the nuclei in tumor cells, whereas adjacent normal colonic epithelium, lymphocytes, and stromal cells had nuclear staining by contrast.
Evaluation of TIL TIL were scored blindly and independently by two observers (Xiaoli Jia and Bin Li). Density of TIL for these tumors as scored using H&E stained tissue sections has been reported previously [8,12]. Tumors were classi ed into TIL-low (< 2 TILs per hpf) and TIL-high (≥ 2 TILs per hpf) cases. For tumors near the cut-point showing a heterogeneous distribution of TILs, scoring was commenced in a hot-spot region identi ed during low power screening, followed by random elds.

Statistical analyses
We adopted Fisher's exact test to compare the correlations among clinicopathological variables, histological subtypes and gene mutations. PFS was de ned as time from surgery to the rst con rmed relapse, with censoring done when a patient died or was alive without recurrence at last contact. Survival analysis was performed with Kaplan-Meier method and survival curves were compared using the log rank test. Cox proportional hazards model was used for univariate and multivariate analysis. Statistical analyses were performed using the SPSS software22.0. P values less than 0.05 were considered statistically signi cant.

Patient characteristics
We identi ed 50 patients with MAC that ful lled the inclusion criteria. There were 19 women and 31 men in this group, ranging in age at diagnosis from 31-90 years of age (median, 67 years). Of the 50 cases, 17 (34 %) involved the right colon, 26 (52 %) the left colon and 7 the rectum (14 %). The average tumor size was 6 cm (range 3-13 cm). The number of patients with pathological stage I, II, III and IV cancer were 3, 20, 23 and 4, respectively (Table 1) Table 1). Unfortunately, there were no statistical signi cance among these groups.

Tumour TIL analysis
In our cohort, TIL status data were available for 50 patients, 28 were classi ed as TIL-high and 22 as TILlow respectively. It is interesting to note that men had lower TIL compared with women (P = 0.018). TILlow tumors were also correlated with advanced MAC stage (P = 0.041) (Fig. 2a). All four IV stage patients were classi ed into TIL-low cases ( Table 2).

Clinico-molecular characteristics
Tumor samples were analyzed for MMR status and somatic mutations in KRAS and BRAF. Five patients (10 %) showed high-frequency MSI (MSI-H) tumors, and 45 patients (90 %) had low frequency MSI (MSI-L) or microsatellite stable (MSS) tumors (2 MSI-L and 43 MSS). There were more early stage tumors (stage I-II) in the MSI-H group (P = 0.048). The remaining clinicopathological factors variables did not signi cantly differ between the groups. We additionally detected 18 (36 %) KRAS mutations. Among these, 10 tumors with G12D mutation, 3 with G12V on exon2, 1 with Q61L and 1 with Q61R mutation on exon3, 1 with A146T and 2 with A146V mutation on exon4. BRAF V600E mutation was found in 4 of 50 (8 %). All of these molecular alterations were mutually exclusive. Nonetheless, there was no statistically correlation between KRAS or BRAF mutation and clinicopathological factors.

Comparative analysis of patient outcome
All patients underwent surgical resection.The follow-up period for PFS were available for 50 patients. 14 % (7 of 50) of individuals experienced a disease recurrence. 7 patients developed metastatic disease at more than one site. Liver and lung metastasis were most frequent, uncommon sites of metastatic diseases were bone and pancreas. It is well known that patients with CRC at stage I have a favorable rate of survival, whereas patients with CRC at stage IV show a poor rate of survival. However, it is di cult to predict the survival of patients with stage II or stage III CRC. Therefore, we analyzed the prognostic value of 43 patients with stage II and III MAC. In the univariate analysis, only TIL-low had a negative association with PFS (P = 0.000). In contrast, patient sex, age, tumor size, tumour location, lymph node metasis, gene mutation, MSI result, lymphovascular or perineurial invasion and MAC histological subtypes showed no signi cant correlation. ( Figure 1H). P53 was expressed in 40 80 % cases and negative in 10 20 % cases. The prognostic value of P53 negative expression exhibited a tendency toward signi cance (P = 0.109) ( Table 3) (Fig. 2b). Nevertheless, multivariate analysis revealed that none of the parameters showed a statistically signi cant association with PFS.
For this reason, this entity is typically underrepresented in CRC research, and a lot of its biology remains poorly understood. Incidence in our department, MAC comprised about 12% (50/400) of surgical resected cases. This study demonstrated that a high frequency of KRAS mutation, but low BRAF and MMR de ciency rates. Among these,the KRAS mutation rate was equal to 38.5%, which was close to most of the previous reports either about Chinese or other ethnicities in CRC patients [2,14.15]. BRAF mutation rates are signi cantly different in previous studies of Chinese population [16]. Like in other studies, these two gene (KRAS and BRAF) mutations were mutually exclusive [17,18]. Similar to Xiaodong Li's study [16], we found that both BRAF or KRAS mutations has no correlation with amount of mucinous component in the tumours.
In the present study, the prevalence of MSI was lower than that reported in the literature [19][20][21]. Our ndings, together with Kepil et al reports, suggest that this nding might have arisen from a geographical/ ethnic difference [22]. Besides, our results did not demonstrate a signi cant relationship between PFS outcome and MMR de ciency. Also, contrary to previous reports [23], we found that P53 negative expression patients were more prone to metastatic relapse, although no signi cant association was found. This may be partially because the number of MMR de ciency or P53 negative expression case is limited in our study.
The mucinous component of CRC has been drawing attention in recent years. Most currently studies comparing CRC with MAC use this criterion, nevertheless, their impact on patients' prognosis has remained controversial. The present results do not agree with some previous reports [3,10] but support those of a recent study by Gonzalez et al [5] and Chen et al [2], in which they state that the proportion of mucinous component of MAC has no signi cant in uence on patient prognosis. While mucinous features should still be mentioned, utilizing criteria such as a 50% cut off is not supported by our data. However, we found the bigger the tumor size, the higher the number of mucinous.
In addition to investigating the prognostic signi cance of the percentage of tumour glands producing mucin, we also evaluated the morphology of the different tumour cells types in mucin pools. MAC tumour cells exhibit highly diverse morphological features, such as layers, acinar or cribriform structures, or individual signet ring cells. SRC is de ned by the presence of ≥ 50% of tumour cells with intracytoplasmic mucin. It has been reported that compared with MAC, patients with SRC were more frequently associated with metastatic disease, metastases at multiple sites, and had a poorer survival than MAC patients [24,25]. Interestingly, during the follow-up period of 32 months, one SRCC patient did not experience any tumor recurrence. The morphology of SRCC case is shown in (Figure. 1c). This patient showed high TIL counts. Our ndings suggest that, to the contrary, no correlation was found between the different morphological types and PFS for the SRC or MAC. Additionallly, while one study did nd that PDCs grade had the highest accuracy in predicting RFS in CRC patients, no signi cant results were found for MAC patients. Taken together, we found histologic characteristics of MAC, including percentage of tumor volume comprised of mucus, were not predictive of outcome.
Lymphocytic in ltration is a major immunological defense against tumor cells in solid tumors and is a potential predictor of CRC [26,27]. Williams et al 8 found TIL status to be a strong independent predictor of PFS in mucinous component tumors, and a superior predictor of prognosis compared with histological grade. The more TIL in ltrates in the tumor microenvironment, the better the prognosis may be. Likewise, we also demonstrated that TIL status was a strong independent predictor of PFS in MAC. Although the histologic subtype of SRCC was once thought to convey additional risk of mortality, our study demonstrates that, SRCC patients with high TIL levels also showed better PFS prognosis. Furthermore, our study found more male patients diagnosed with TIL-low tumors than female patients. The results of our study suggest that using the grade of TIL status instead of WHO grade of the entire tumor can help identify patients with a high risk of recurrence more accurately.
This study still had certain limitations. Firstly, this study was a retrospective analysis, and these patients' long-term survival could not be obtained at present. Secondly, limitations of our study include its small patient sample, as well as the low number of adverse events (recurrences), both factors likely contributed to the lack of statistical signi cance of TIL status in multivariate analysis. A larger sample size may be needed to adequately detect these issues in the future.
In summary, TIL status showed statistically signi cant association with PFS. In contrast, other parameters, including KRAS or BRAF mutational status MMR de ciency percent of glands producing mucin and the morphology of the different tumor cell types in mucin pools showed no signi cant correlation. Consequently, TIL assessment should be considered for introduction into clinical practice and included in the pathology report of MAC. The result also revealed that the current cut off of 50% mucin component to de ne MAC might be challengeable. To con rm these ndings, further studies with larger sample size are needed. Availability of data and materials All data generated or analyzed during this study are included in the submission. The raw data are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
Institutional and/or national research ethic committee has approved the data collection and management process.