Evaluation of the Relationship Between Portal Vein Diameter and Colorectal Liver Metastases on Computed Tomography

Discussion

CRC is a common neoplasm in western countries, mostly due to nutrition, and its incidence is increasing in developing countries due to improvements in food and lifestyle (14). The most common and often first metastatic site of CRC is the liver. Approximately 25% of patients present with liver metastasis at the time of initial diagnosis, and 50% of patients develop liver metastasis during the course of CRC (14, 15). Appropriate patient selection is also necessary to achieve oncological benefits in the curative resection of liver metastases. Radiological evaluation of CRLM is pivotal in determining treatment strategies, including decisions regarding preoperative surgery and chemotherapy (16). Preoperative chemotherapy is also becoming increasingly useful in shrinking initially unresectable lesions, thereby making surgery feasible (17). USG, CT, MRI, and PET are the main modalities used to detect CRC and its metastases. Imaging is also routinely used for local and distant tumor staging (18, 19).

CRLM usually presents with a heterogeneous appearance, indicating the presence of fibrosis, necrosis, and tumor cells spreading from the center of the tumor. The metastasis content tends to become homogeneously hypoattenuated as a result of infarct-like necrosis (ILN) that develops after chemotherapy (9, 20). It is also thought that the heterogeneous appearance of CRLM lesions reflects the predominance of usual necrosis (UN). When an abdominal CT scan with IV is performed, CRLMs usually appear as faint and blurry lesions with irregular borders. However, when chemotherapy is administered, necrosis develops within the tumor and the contrast difference with normal liver parenchyma becomes apparent, making the borders of the tumor easier to recognize (21). Different imaging findings have also been reported on CT, such as desmoplastic reaction, inflammatory cell infiltration, and residual tumor proliferation, corresponding to the IV effect at the tumor margins (3, 21, 22). CRLM has the same histological features as the primary tumor and is often described as “not otherwise specified” (NOS) adenocarcinomas. Mucinous tumor is the most common of the rare subtypes (2, 23).

Synchronous and metachronous LMs present additional challenges in patients with CRC, and previous studies have been conducted on this issue. Okholm et al. evaluated the incidence and prognosis of LMs in patients with CRC. A total of 1,672 patients were included and 23.6% were diagnosed with CRLM. Synchronous metastasis was present in 16% of the cases and metachronous metastasis was present in 7.7%. The incidence of synchronous metastases has been found to be high despite diagnostic innovations. While treatment options for CRLM have improved, it has been noted that patient survival is significantly reduced in the presence of metastatic disease (4). Another study was based on the idea that there is no consensus on the defining time point for synchronous/meta-synchronous and its prognostic implications remain unclear. Engstrand et al. attempted to determine its prognostic value for differential diagnosis by making measurements at various time points in a population-based patient cohort. A total of 1,026 patients diagnosed with CRC in Stockholm and Gotland were included in the study and followed for five years or until death to identify patients diagnosed with CRLM. The cases with CRLM were followed for at least five years from the time of diagnosis of metastases or until death. CRLM was detected in 272 patients, and there was no statistically significant difference in overall survival between synchronous and metachronous detection at any of the defining time points (first diagnosis/surgery and 3, 6, and 12 months). The results obtained were compared with 41 similar publications in the literature, but a clear model regarding the prognostic importance of synchronous and metachronous detection could not be revealed (10).

Another important area of study is the diagnosis of CRLM using radiological methods. Tamura et al. evaluated peripheral enhancement on IV abdomen tomography after chemotherapy in patients with CRLM in terms of its relationship with corresponding pathological findings. As a result of archive scanning, 44 CRLM cases who underwent hepatic resection after preoperative chemotherapy between 2008 and 2013 were evaluated and divided into three groups according to the contrast enhancement pattern (from smooth edge to fuzzy margin). The percentage of infarct-like necrosis was significantly higher in patients with CRLM with smooth borders than in those with blurry edges, and the number of viable cells was lowest in this group (p<0.001). Accordingly, it was stated that the type of necrosis was related to the nature of the edges and the presence of residual cells was related to peripheral contrast enhancement (3). Similarly, Ishida et al. investigated the relationship between the morphological appearance on CT and the histological findings of CRLM after preoperative chemotherapy. Forty-seven patients who underwent first hepatic resection for CRLM after preoperative chemotherapy and had available contrast-enhanced CT scans were examined. Morphological appearances on CT included metastases ranging from heterogeneous to mixed and homogeneous lesions, tumor-liver interface, and peripheral edge enhancement on CT. Histological parameters mainly included UN, ILN, three-zone change, dangerous halo, and mucosal lake cells. The widespread attenuation of the contrast agent on CT revealed that UN predominance was more common in the heterogeneous group than in the mixed and homogeneous groups (p<0.05). In the ILN-dominant group, on the contrary, the decrease in attenuation was greater in the homogeneous group (p<0.05). The histological presence of live tumor cells was also closely related to the tumor-liver interface and peripheral margin increase (p<0.01). As a result, it was shown that the morphological findings on CT regarding necrosis were related to the histological findings (9).

Choi et al. attempted to systematically determine the diagnostic accuracy of multidetector CT, contrast-enhanced MRI, and PET/CT for the diagnosis of CRLM and the sources of heterogeneity among reported results. In this meta-analysis, 2,151 lesions from CT studies, 2,301 lesions from MRI studies, and 1,846 lesions from PET/CT studies in the literature were evaluated. Although neoadjuvant chemotherapy did not significantly affect the sensitivity of PET/CT, it did reduce the sensitivity of CT and MRI (p<0.01). Despite the heterogeneous accuracy between studies, MRI with IV showed the highest sensitivity for all cases, and MRI and PET/CT had similar specificities in diagnosing CRLM (12). In another meta-analysis, Niekel et al. compared the diagnostic performance values of CT, MRI, PET, and PET/CT for the detection of CRLM in cases without prior treatment. Articles containing at least 10 patients with histopathologically proven CRC from January 1990 to January 2010 were searched. A total of 3,391 patients were identified and these four imaging modalities were evaluated. It was suggested that MRI could serve as the primary method and FDG PET as a secondary approach for evaluating CRLMs (13).

Previous studies on CRLM have mostly focused on evaluating the enhancement pattern of lesions and the diagnostic effectiveness of different imaging modalities (3, 9, 12, 13). Additionally, some work has been done on synchronous and metachronous CRLMs (4, 10). Unlike the previous ones, in this study, a completely different method was chosen and, for the first time in the literature, the possible relationship between the diameters of the portal veins and the diameters of the CRLMs was evaluated. Although no significant relationship was found between portal vein diameters and CRLM diameters in our study (p>0.05), it is noteworthy that there was a significant association between the presence of metastasis in the left lobe of the liver and the metastasis diameter (p<0.05). Perhaps our findings suggest that CRLMs located in the left liver lobe may be more amenable to preoperative chemotherapy regimens and surgical resectability (14, 15, 18, 19). Another advantage of our study is that, unlike many other studies, the cases in this study did not receive any treatment, including preoperative chemotherapy. Thus, it allowed us to evaluate the relationship between portal veins and CRLM diameters more objectively (3, 4, 9-13, 21).

Study limitations. The first limitation is the lack of similar studies in the literature with which we can compare our results. It is obvious that there is a need for new studies with higher number of cases to enable comparisons similar to our measurements. The second limitation is the relatively low number of cases in group b (two metastases), with 23 cases. The third limitation is that the number of cases originating from right colon tumors (42 cases) was lower than those originating from the left colon (133 cases). The fourth limitation is that due to the retrospective nature of the study, there may inevitably be selection bias. Another limitation is that, unlike in other studies, CRLMs were evaluated only with CT imaging (12, 13).