Histopathological prediction of liver metastasis after curative resection of colorectal cancer

Local massive venous invasion in colorectal cancer: CT-pathological correlation and its clinical implication

OBJECTIVE

To identify CT findings of massive venous invasion (MVI) in colorectal cancer, compare them to pathological findings and evaluate its clinical implications.

METHODS

Among 423 patients who received surgical resection of colorectal cancer, pre-operative CT of 26 patients (15 males, 11 females; mean age, 63.0 ± 12.1 years) with histopathologially proven MVI and 26 patients (14 males, 12 females; mean age, 71.1 ± 9.6 years) with histopathologically proven lymph node (LN) metastases were reviewed and compared with histopathological findings. We evaluated CT detectability of MVI and the morphologic differences between MVI and LN metastasis. All cases were followed up for at least 6 months after surgery.

RESULTS

Pre-operative CT correctly diagnosed only one case as tumour thrombus. 9 lesions were not detected on CT, and others were misdiagnosed pre-operatively as regional LN metastasis (14 cases) and juxtatumoural abscess (2 cases). After reviewing these cases, MVIs were identifiable in 20 of 26 cases. MVI was depicted on CT as nodules (oval, lobulated), abscess-like or intravenous tumour thrombus. MVI was significantly larger than LN metastasis (p < 0.05), while contrast enhancement was significantly lower (p < 0.05), and MVI often had an enhanced rim. Ten patients had synchronous metastases, and six patients had metachronous distant metastases within 5 years.

CONCLUSION

Many cases of MVI were distinguishable from LN metastases on pre-operative CT of colorectal cancer, but their appearances were varied, reflecting their histopathological behaviours. The distant metastatic rate was much higher in cases with MVI.

ADVANCES IN KNOWLEDGE

Radiologists should be aware of CT findings of MVI in colorectal cancer as a sentinel sign of distant metastasis.

Preoperative FDG-PET predicts early recurrence and a poor prognosis after resection of pancreatic adenocarcinoma

OBJECTIVE

The aim of this study was to evaluate the clinical utility of preoperative 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) as a prognostic predictor of resectable pancreatic adenocarcinoma.

METHODS

A total of 128 patients with resected pancreatic adenocarcinoma who preoperatively underwent FDG-PET examinations were studied. The maximum standard uptake value (SUVmax) was calculated for each primary lesion. A receiver operating characteristics (ROC) curve was constructed to estimate the optimal cutoff value of the SUVmax. In order to determine which outcomes of interest were appropriately demonstrated, ROC curve analyses were conducted for six outcomes: 6-, 12-, and 24-month disease-free survival (DFS), and 6-, 12-, and 24-month overall survival (OS). A multivariate analysis was conducted to identify prognostic factors.

RESULTS

The ROC curves revealed that the SUVmax predicted the 6-month DFS most optimally (area under the curve 0.757), with a cutoff value of 6.0. Of the 69 patients with an SUVmax ≥ 6.0, 34 (49 %) developed recurrence within 6 months. In contrast, only 3 of 59 (5 %) patients with an SUVmax < 6.0 exhibited early recurrence (p < 0.001). The median OS time was 37 months in patients with an SUVmax < 6.0 and 18 months in patients with an SUVmax ≥ 6.0 (p < 0.001). The multivariate analysis revealed lymph node metastasis (hazard ratio [HR] 2.31; p = 0.001) and an SUVmax ≥ 6.0 (HR 2.05; p = 0.002) to be significantly correlated with a poor survival.

CONCLUSIONS

An SUVmax ≥ 6.0 was a significant predictor of early postoperative recurrence and subsequent poor survival following resection of pancreatic adenocarcinoma.

Combination of haptoglobin and osteopontin could predict colorectal cancer hepatic metastasis

PURPOSE

This study aimed to identify novel molecular markers for the early detection of colorectal cancer liver metastasis.

METHODS

Genes related to hepatic metastasis were screened from the Oncomine database. The candidate markers were tested by immunohistochemistry, and their predictive accuracy was assessed by the cross-validation method and an independent test set.

RESULTS

We got three datasets containing 2,973 genes that were highly expressed in primary colon cancer tissues compared with non-metastatic colon cancer tissues and identified 7 candidate molecules for immunohistochemical validation. A total of 213 colorectal cancer samples were randomly divided into a training set (113 cases) and a blind testing set (100 cases). In the training set, immunohistochemical analysis showed that HP, OPN, and PTGIS expression were significantly higher in the hepatic metastasis group than in the non-metastasis group. Logistic regression analysis showed that HP and OPN levels in primary tumors and lymph node metastasis status were the only significant (P<0.05) parameters for detecting liver metastasis. The predictive accuracy of markers was assessed by the cross-validation method and an independent test set. In leave-one-out cross-validation, the two markers combined with clinicopathologic features resulted in 91.2% sensitivity and 96.4% specificity for hepatic metastasis detection. In an independent test set, the combination achieved 94.5% sensitivity and 88.9% specificity for predicting the hepatic metastasis of colorectal cancer.

CONCLUSIONS

Our results suggest that combined HP and OPN expression levels are significantly related to liver metastasis and prognosis, and, if this is validated, they could be used as accurate predictors of liver metastasis in colorectal cancer.

Nuclear beta-catenin overexpression in metastatic sentinel lymph node is associated with synchronous liver metastasis in colorectal cancer

BACKGROUND

Beta-catenin, a component of the Wingless/Wnt signaling pathway, can activate target genes linking with the adenomatous polyposis coli (APC) gene in colorectal cancer. The purpose of this study is to investigate whether nuclear beta-catenin overexpression in metastatic sentinel lymph node(s) [SLN(s)] is associated with synchronous liver metastasis.

METHODS

Clinicopathological data from 355 patients (93 cases with liver metastasis and 262 cases without liver metastasis) were reviewed. Beta-catenin expression in metastatic SLN(s) and liver metastatic lesions was examined by immunohistochemistry. The association of nuclear beta-catenin expression in metastatic SLN(s) and liver metastatic lesions was evaluated, and the relationship between nuclear beta-catenin expression and clinicopathological characteristics was analyzed. Finally, univariate and logistic multivariate regression analyses were adopted to discriminate the risk factors of liver metastasis.

RESULTS

Nuclear beta-catenin overexpression in metastatic SLN(s) was observed in 70 patients with liver metastasis and 31 patients without liver metastasis (75.3% vs. 11.8%; P < 0.001). Nuclear beta-catenin expression was noted in all the metastatic lesions. Spearman rank correlation analysis demonstrated that nuclear beta-catenin expression in metastatic SLN(s) had a positive correlation with that in metastatic lesions (r = 0.425, P < 0.001). Univariate and multivariate analyses indicated that nuclear beta-catenin overexpression in metastatic SLN(s) correlated with liver metastasis.

CONCLUSIONS

Nuclear beta-catenin overexpression in metastatic SLN(s) is strongly associated with liver metastasis and may contribute to predict liver metastasis.

Prognostic value of nuclear β-catenin overexpression at invasive front in colorectal cancer for synchronous liver metastasis

BACKGROUND

β-catenin plays an important role in colorectal tumorigenesis. Relatively little is known about the relationship between β-catenin overexpression and liver metastasis. The purpose of this study was to investigate whether nuclear β-catenin overexpression in colorectal cancer is associated with synchronous liver metastasis.

METHODS

The β-catenin expression in tumor tissue from 486 patients with colorectal cancer was examined by immunohistochemistry. The relationship between nuclear β-catenin expression in colorectal cancers and liver metastatic lesions and other clinicopathological characteristics was analyzed. Univariate analysis and logistic multivariate regression analysis were adopted to discriminate risk factors of liver metastasis.

RESULTS

Nuclear β-catenin overexpression at the invasive front of the primary tumor in patients with liver metastasis is more evident than that in patients without liver metastasis (71.5% vs. 29.3%; P < 0.001). Nuclear β-catenin expression in primary tumors had a positive correlation with that in the matched metastatic lesions (r = 0.499, P < 0.001). Univariate and multivariate analyses indicated that overexpression of nuclear β-catenin at the invasive front in colorectal cancer correlated with liver metastasis.

CONCLUSIONS

Overexpression of nuclear β-catenin at the invasive front in colorectal cancer is strongly associated with liver metastasis and may be a promising predictor of liver metastasis.

Association between age and synchronous liver metastasis in female colorectal cancer patients

PURPOSE

The purpose of this study is to investigate the relationship between age and synchronous liver metastasis in female colorectal cancer patients.

METHOD

Clinical and pathological data from 655 consecutive female patients who were treated in Qilu Hospital from January 2000 to January 2010 were reviewed. First, the patients were divided into two groups: 60 years or younger and 61 years or older. A chi-square test was adopted to analyse the difference in clinicopathological characteristics between the two groups. Then, they were classified into two groups according to liver metastasis. Univariate analysis and logistic multivariate regression analysis were adopted to discriminate risk factors of liver metastasis.

RESULTS

The chi-square test demonstrated that significant difference existed between the younger and the older groups in terms of liver metastasis (P < 0.001), lymph node metastasis (P < 0.001), tumour localization (P < 0.001), tumour invasion depth (P < 0.001), type of tumour (P < 0.001), tumour cell differentiation (P < 0.001) and venous invasion (P < 0.001). Univariate analysis demonstrated that seven factors are associated with liver metastasis. Logistic regression analysis indicated that age (P = 0.005), tumour size (P < 0.001), tumour invasion depth (P = 0.001), tumour cell differentiation (P = 0.029) and type of tumour (P < 0.001) are independent risk factors of liver metastasis.

CONCLUSIONS

The liver metastatic potential of colorectal cancer may be different between younger and older female patients. Age may independently influence liver metastasis in female colorectal cancer patients.

Quantitative measurement of venous invasion of colorectal cancer with metachronous liver metastasis

AIMS

Many studies have proven the importance of venous invasion in colorectal cancer with synchronous liver metastasis. The aim was to clarify the relationship between venous invasion and metachronous liver metastasis, which is not fully understood.

METHODS AND RESULTS

A histological study of venous invasion in colorectal carcinoma was performed using a total of 156 patients, of whom 52 survived without recurrence for 5 years (Group A); 47 had metachronous liver metastasis (Group B), and 57 had synchronous liver metastasis (Group C). The number and the maximum area of venous invasion were estimated in each case per x 40 field of cancerous lesions, which were divided into intramural and extramural lesions. A high incidence and high average number of foci of venous invasion appeared in Groups B and C. The average maximum areas of extramural venous invasion were much larger in Groups B and C than in Group A.

CONCLUSIONS

The average number of foci of venous invasion by colorectal cancer with metachronous liver metastasis did not differ significantly from that with synchronous ones. Furthermore, invasion into extramural large veins appeared to be associated with liver metastasis.

Biology of colorectal liver metastases: A review

Metastatic growth is a selective, non-random process, which in the case of colorectal cancer, frequently occurs in the liver and is the major cause of cancer related death in these patients. This review summarises attempts to find biological and molecular markers of metastasis and their role in establishment of secondary tumours. Recent evidence suggests that liver metastases are phenotypically different to the primary from which they were derived and thus represent a separate disease entity.

Highly proliferative intratumoral fibroblasts and a high proliferative microvessel index are significant predictors of tumor metastasis in T3 ulcerative-type colorectal cancer

Fibroblast and endothelial cell mitotic figures are seen in some areas of colorectal cancers, and the purpose of this study was to investigate whether the proliferative activity of fibroblasts and endothelial cells plays an important role in tumor progression of T3 ulcerative-type colorectal cancer. The tumor area of 157 colorectal cancers was divided into marginal elevated area and central depressed area (CDA), and at half the depth of the depression the CDA was in turn divided into CDA upper area (CDAU) and CDA lower area (CDAL). The proliferative activity of the tumor cells, fibroblasts, and endothelial cells was assessed immunohistochemically by double CD31/MIB-1 (anti--Ki-67 antigen) staining. The proliferative microvessel index was estimated as the percentage of microvessels lined by MIB-1-positive endothelial cells relative to the total microvessel count. Proliferative activities of tumor cells showed significant associations with those of fibroblasts and the proliferative microvessel indices in all of the corresponding areas. Proliferative activities of fibroblasts also showed significant associations with proliferative microvessel indices in all of the corresponding areas. Colorectal cancers with nodal metastasis showed significantly higher proliferative activities of fibroblasts in the CDAU than those without nodal metastasis (P <.001). The high proliferative activities of fibroblasts and proliferative microvessel indices in the CDAU showed significant associations with short distant organ metastasis-free periods in colorectal cancers without nodal metastasis (P <.001 and P =.010, respectively) and those with nodal metastasis (P =.024 and P =.036, respectively). Multivariate analysis showed that the highly proliferative fibroblasts in the CDAU significantly increased hazard rates of distant organ metastasis of colorectal cancer patients with nodal metastasis (P =.018). Proliferative activities of fibroblasts and endothelial cells in the CDAU are useful parameters for predicting tumor metastasis in patients with T3 ulcerative-type colorectal cancer. HUM PATHOL 32:401-409.

Tumor thickness is a histopathologic predictive parameter of tumor metastasis and prognosis in patients with Dukes stage C ulcerative-type colorectal carcinoma. A two-hospital-based study

BACKGROUND

Metastasis to the liver or lymph nodes is an important prognostic factor in patients with colorectal carcinoma. The purpose of the current study was to estimate the power of tumor thickness in predicting metachronous liver metastasis (MLM), lymph node metastasis (LNM), or overall survival (OS) in patients at two hospitals (the National Cancer Center Hospital [NCCH] and the National Cancer Center Hospital East [NCCHE]) to confirm the reproducibility of the study.

METHODS

The subjects of this study were 74 and 186 consecutive patients with ulcerative-type colorectal carcinoma treated at the NCCH and NCCHE, respectively. Tumor thickness was measured in three areas: 1) the marginal elevated area (MEA), 2) the central depressed area (CDA), and 3) the most thickened area (MTA). Studies were performed with well known histologic parameters to compare the predictive power of tumor thickness on MLM, LNM, and OS using the Cox proportional hazards regression model or analysis of variance.

RESULTS

A significant correlation between tumor thickness and MLM was observed only in the CDA in the NCCH patients (P = 0.005). The authors applied a tumor thickness cutoff value in the CDA of 10 mm (</= 10 mm and > 10 mm) for further study. Multivariate analyses demonstrated that a tumor CDA thickness > 10 mm was associated significantly with MLM, multiple LNMs, and OS in NCCH patients with Dukes Stage C disease (P = 0.002, P = 0.023, and P = 0.002, respectively). A significant predictive power for tumor CDA thickness for MLM, multiple LNMs, and OS was confirmed by multivariate analysis in NCCHE patients with Dukes Stage C disease (P = 0.008, P = 0.021, and P = 0.010, respectively).

CONCLUSIONS

The CDA thickness of the tumor was found to be a useful predictive parameter for MLM, multiple LNMs, and OS in patients with Dukes Stage C ulcerative-type colorectal carcinoma who were being treated in two independent hospitals.

A comparison of staging systems for localized extremity soft tissue sarcoma

BACKGROUND

Staging systems for soft tissue sarcoma (STS) are important to identify patients with similar systemic risk who might benefit from specific treatments. This study compared four commonly used staging systems for predicting systemic outcomes of patients with localized extremity STS, as proposed by the fourth and fifth editions of the American Joint Committee on Cancer/International Union Against Cancer (AJCC/UICC) staging system, the Memorial Sloan-Kettering Cancer Center (MSK) system, and the Surgical Staging System (SSS) of the Musculoskeletal Tumor Society.

METHODS

Three hundred consecutive adult patients with newly diagnosed nonmetastatic STS of the lower extremity were treated at Memorial Sloan-Kettering Cancer Center between 1982 and 1989. Metastasis free survival was the end point of the study. The prognostic value of the four staging systems and their components were examined in univariate and multivariate analyses. The Akaike information criterion (AIC) was used to identify the system that best predicted the risk of systemic recurrence.

RESULTS

Compartment status, depth, grade, and size were all independent predictors of outcome within their respective staging systems. However, when compared with one another, only depth, grade, and size retained their prognostic significance. Of the four models, the AIC predicted that the MSK was the best predictor of systemic relapse, followed by the fifth edition of the AJCC/UICC staging system.

CONCLUSIONS

Staging systems such as the MSK system or the fifth edition of the AJCC/UICC system, which include tumor depth, grade, and size as prognostic factors, are the most predictive of systemic relapse in patients presenting with localized extremity STS. Both of these systems identify the same group of patients at the highest risk who would be the most suitable for adjuvant chemotherapy trials.

Macroscopic features at the deepest site of tumor penetration predicting liver metastases of colorectal cancer

Liver metastasis is the gravest prognostic factor in colorectal cancer. To identify a reliable indicator for liver metastasis, we evaluated macroscopic features and seven established histopathological findings at the cut section containing the deepest penetration using univariate and multivariate analyses in 417 colorectal cancers. Macroscopic features were divided into two types, streak type and non-streak type, according to the presence or absence of white streak(s) at the advancing margin of tumor invasion. Streak type was observed in 109 patients (26%). The frequency of liver metastasis in streak type tumors (56%) was significantly higher than that in non-streak type tumors (13%) (p < 0.001). The white streak corresponded histologically with cancer cells showing focal dedifferentiation with marked stromal and perivascular fibrosis extending towards the serosa or adventitia. In 343 curatively treated patients, univariate analysis showed that recurrent liver metastasis was significantly associated with macroscopic features, venous invasion, focal dedifferentiation and lymph node metastasis. Multivariate analysis disclosed that macroscopic features and lymph node metastasis were independent indicators of liver metastasis. These macroscopic features, corresponding histologically to stromal behavior against invading cancer cells, are a simple and useful indicator of liver metastasis of colorectal cancer.

Prognostic value of pathological characteristics of colorectal cancer

The overall cure rate of colon cancer has not improved dramatically in the last decade, remaining at approximately 60% 5-year survival. The main reason for this lack of progress is that at the moment the primary tumour is resected, a significant proportion of the patients with seemingly localised disease already has (undetectable) micrometastases, mostly in the liver. The most important prognostic indicators have been extension of the tumour into the bowel wall and the presence of lymph node metastasis, as expressed in the Dukes classification. However, in the Dukes B and C categories, these parameters are poor predictors of final outcome. For improvement of the prognosis, in addition to earlier detection, more aggressive (adjuvant) treatment of high risk patients would be a rational strategy. This requires development of new therapeutic modalities, but also reliable stratification of patients according to high risk or low risk for recurrent disease. In recent years, many attempts have been made to improve the prediction of final outcome. Parameters studied include inflammatory response to the primary tumour, tumour cell growth fraction, tumour cell differentiation, genetic abnormalities and expression of genes involved in invasion and metastasis. Although some of these newer parameters have significant predictive value, in multivariant analyses, most appear to have limited independent value. Recent studies indicate that genetic abnormalities might be important new prognostic indicators. One of the most promising findings in this area is an allelic loss of chromosome 18q, which allows division of Dukes B patients into subgroups with low risk and high risk for recurrent disease.