Dual Stain With SATB2 and CK20/Villin Is Useful to Distinguish Colorectal Carcinomas From Other Tumors

Dual Stain With SATB2 and CK20/Villin Is Useful to Distinguish Colorectal Carcinomas From Other... Abstract Objectives Small sample size limits the number of immunostains that may be attempted in colorectal carcinoma (CRC) biopsy specimens. We investigated the utility of dual stain with special AT-rich sequence binding protein 2 (SATB2) or caudal-type homeobox 2 (CDX2) and cytokeratin 20 (CK20) or villin in identifying CRC. Methods Tissue microarrays with 222 CRCs and 375 other carcinomas were built. Dual stain was performed pairing nuclear stains CDX2 or SATB2 with CK20 or villin. Results All four single stains showed excellent sensitivity (93%-99%) but variable specificity (56%-88%) for CRC. All four dual stains also showed excellent sensitivity (90%-96%) while much improved specificity (88%-98%) compared with single stains. SATB2 dual stain (with CK20 or villin) showed a higher specificity than CDX2 dual stain (with CK20 or villin) with a comparable sensitivity. Conclusions SATB2 dual stain shows the greatest potential clinical utility in identifying CRC and is superior to CDX2 dual stain. More important, SATB2 dual stain could be helpful for specimens with limited tissues or those having a nonclassic staining pattern. Dual immunostain, SATB2, CDX2, CK20, Villin, Colorectal tumor Colorectal carcinoma (CRC) is one of the most common cancers worldwide and in the United States.1 A significant proportion of patients with CRC initially present with distant metastases, frequently requiring core biopsy for diagnosis and to direct further management. In such case, immunohistochemistry (IHC) is usually performed to identify the primary site. Cytokeratin 20 (CK20) is an intermediate filament protein expressed in gastrointestinal (GI) tract glandular epithelial cells and is routinely used as a diagnostic marker for CRC.2-4 CK20 is a highly sensitive marker for CRC. However, its specificity is low due to its expression in several other types of adenocarcinomas.2-4 Similar results have also been found with villin, another marker that has been used in the GI tract. Caudal-type homeobox 2 (CDX2) is a transcription factor regulating intestinal epithelial cell differentiation that is strongly expressed in up to 100% of CRCs but also expressed in gastric carcinomas (18%) and ovarian mucinous tumors (20%).5-10 Special AT-rich sequence binding protein 2 (SATB2) is a DNA-binding protein expressed in glandular cells lining the lower GI tract.11 SATB2 is a highly sensitive and also specific marker for CRC, and SATB2 in combination with CK20 can identify 97% of CRCs.12 The same study also revealed that only 3.3% of ovarian carcinomas and 5.7% of lung adenocarcinomas were positive for SATB2, while all gastric and pancreatic carcinomas were all negative for SATB2.12 Additional study has demonstrated that SATB2 showed a high sensitivity (93%) and specificity (77%) to identify a metastasis as having a colorectal origin, and in combination with CK7 and CK20, the specificity increased to 100%.13 Decreased SATB2 expression level has been shown to associate with metastasis and poor prognosis in CRCs.14,15 SATB2 is also a marker of osteoblastic differentiation in benign and malignant mesenchymal tumors.16,17 Metastatic colorectal carcinoma biopsy specimens sometimes have limited materials for further workup; therefore, dual stains or multiplex immunostains are very efficient to use limited materials. Given the high sensitivity and specificity of SATB2 in CRC and the challenge in identifying the primary site for metastatic CRC, we were interested in its utility in a dual stain with either CK20 or villin compared with CDX2 in a dual stain with the same two markers in identifying CRC on limited materials. In this study, we explored dual stains with one of the nuclear stains SATB2 or CDX2 in combination with the cytoplasmic stain of CK20 or villin and compared their sensitivities and specificities to determine whether any combination could provide improved diagnostic utility for CRC. Materials and Methods Case Selection and Microarray Construction After institutional review board approval at Ohio State University (OSU), an electronic pathology archive database search was performed to select CRCs diagnosed from 2000 to 2011 at the OSU Wexner Medical Center. In total, 222 CRCs and 375 other carcinomas (45 breast, 67 hepatocellular, 50 lung, 32 neuroendocrine including 31 from pancreas and one from small bowel, 18 ovarian mucinous tumors, 40 pancreas, 45 prostate, and 78 gastric) were identified, and the diagnosis was confirmed after reviewing pathologic slides and correlating with patients’ clinical history. All cases in our cohort were from the primary site of origin. A tumor block representative of the primary tumor was collected from each case, and tissue microarrays (TMAs) were constructed with duplicate cores of 1 mm for each tumor. Dual Immunostains and Scoring dual stains were performed on TMAs by pairing nuclear stains CDX2 (1:300; BioGenex, San Ramon, CA) or SATB2 (1:50; Santa Cruz, Santa Cruz, CA) with cytoplasmic stains CK20 (1:200; Dako, Carpenteria, CA) or villin (1:50; Cell Marque, Rocklin, CA). The detailed staining protocols were previously described.18 Briefly, 4-μm sections were cut from TMA blocks. CDX2 and SATB2 were labeled as brown color and CK20 and villin were labeled as red color. A Bond Polymer Refine Detection Kit (Leica Biosystems, Buffalo Grove, IL) was used as the brown chromogen, and Bond Polymer Refine Red Detection Kit (Leica Biosystems) was used as the red chromogen. Stained slides were independently reviewed by three pathologists (J.B.R., R.R., and W.L.F.), and a consensus staining result was reached for each case. Staining results were scored as positive or negative with a positive threshold of staining of 5% or more of tumor cells. Representative images are demonstrated in Image 1 and Image 2 . Image 1 View largeDownload slide Representative images of dual stains with SATB2 or CDX2 and CK20 or villin in moderately differentiated and poorly differentiated colorectal carcinomas. SATB2 and CDX2 showed brown nuclear staining, while CK20 and villin showed red cytoplasmic staining (×200). CDX2, caudal-type homeobox 2; CK20, cytokeratin 20; SATB2, special AT-rich sequence binding protein 2. Image 1 View largeDownload slide Representative images of dual stains with SATB2 or CDX2 and CK20 or villin in moderately differentiated and poorly differentiated colorectal carcinomas. SATB2 and CDX2 showed brown nuclear staining, while CK20 and villin showed red cytoplasmic staining (×200). CDX2, caudal-type homeobox 2; CK20, cytokeratin 20; SATB2, special AT-rich sequence binding protein 2. Image 2 View largeDownload slide Representative images of dual stains with CDX2/SATB2 and CK20/villin in different carcinomas. A colonic adenocarcinoma showed negative CDX2 and CK20 staining with CDX2/CK20 dual stain but positive SATB2 staining with SATB2/CK20 dual stain. A lung adenocarcinoma showed positive CDX2 staining with CDX2/CK20 dual stain but negative SATB2 and CK20 staining with SATB2/CK20 dual stain. A pancreatic adenocarcinoma showed positive CDX2 staining with CDX2/villin dual stain but negative SATB2 and villin staining with SATB2/villin dual stain. A stomach adenocarcinoma showed positive CDX2 and CK20 staining with CDX2/CK20 dual stain but negative SATB2 and positive CK20 staining with SATB2/CK20 dual stain (×200). CDX2, caudal-type homeobox 2; CK20, cytokeratin 20; SATB2, special AT-rich sequence binding protein 2. Image 2 View largeDownload slide Representative images of dual stains with CDX2/SATB2 and CK20/villin in different carcinomas. A colonic adenocarcinoma showed negative CDX2 and CK20 staining with CDX2/CK20 dual stain but positive SATB2 staining with SATB2/CK20 dual stain. A lung adenocarcinoma showed positive CDX2 staining with CDX2/CK20 dual stain but negative SATB2 and CK20 staining with SATB2/CK20 dual stain. A pancreatic adenocarcinoma showed positive CDX2 staining with CDX2/villin dual stain but negative SATB2 and villin staining with SATB2/villin dual stain. A stomach adenocarcinoma showed positive CDX2 and CK20 staining with CDX2/CK20 dual stain but negative SATB2 and positive CK20 staining with SATB2/CK20 dual stain (×200). CDX2, caudal-type homeobox 2; CK20, cytokeratin 20; SATB2, special AT-rich sequence binding protein 2. Statistical Analysis The sensitivity, specificity, and overall percentage correct in identifying CRCs were calculated with exact binomial 95% confidence intervals (CIs) for each stain and dual stain separately. Sensitivity and specificity were directly compared between selected stains using the McNemar test for paired comparisons; exact P values were calculated. The 95% CIs for the differences between sensitivities or specificities were calculated using methods based on the score interval as previously described.19,20 All analyses were performed using SAS/STAT software version 9.4 of the SAS System for Windows (SAS Institute, Cary, NC). Results We first evaluated the staining results in 222 CRCs and 375 other tumors for each marker individually. As shown in Table 1 , SATB2 nuclear staining was present in 95% of CRCs, followed by 38% of prostate cancers, 22% of ovarian cancers, 16% of breast cancers, 12% of lung cancers, 7% of hepatocellular carcinomas, 6% of neuroendocrine tumors, 5% of gastric cancers, and 3% of pancreatic cancers. CK20 and villin cytoplasmic staining was present in 93% and 97% of CRCs, respectively, and CDX2 nuclear staining was seen in 99% of CRCs. In our cohort, 197 CRCs were positive for all of the following three stains: CDX2, CK20, and SATB2. Two CRCs (0.9%) were CDX2 negative, and both were positive for SATB2 and CK20. Three (1.4%) CRCs were CK20 negative and STAB2 negative but CDX2 positive. Twelve (5.4%) CRCs were CDX2 positive and SATB2 positive but CK20 negative. Eight (3.6%) CRCs were SATB2 negative but CDX2 positive and CK20 positive. There was no case that was both CDX2 negative and CK20 negative in our cohort. Next we evaluated the dual staining results. Dual stain with SATB2 and CK20 showed positivity for both markers in 90% of CRCs and in 0% to 9% of other tumors. Dual stain with SATB2 and villin showed positivity for both markers in 93% of CRCs and 0% to 7% of other tumors. Although dual stains with CDX2 and CK20/villin showed positivity for both markers in up to 96% of CRCs, they were positive in up to 44% of other tumors (Table 1). Table 1 Frequency of Single Stains and Dual Stains in Different Primary Tumors Type   No.   CDX2 + CK20, No. (%)  CDX2 + Villin, No. (%)  SATB2 + CK20, No. (%)  SATB2 + Villin, No. (%)  CDX2, No. (%)  SATB2, No. (%)  Villin, No. (%)  CK20, No. (%)  Colorectal  222  205 (92)  213 (96)  199 (90)  206 (93)  220 (99)  211 (95)  215 (97)  207 (93)  Breast  45  0  0  0  0  0  7 (16)  2 (4)  0  Hepatocellular  67  0  0  3 (4)  5 (7)  0  5 (7)  34 (51)  9 (13)  Lung  50  2 (4)  6 (12)  0  3 (6)  6 (12)  6 (12)  29 (58)  8 (16)  Neuroendocrine  32  0  3 (9)  0  1 (3)  9 (28)  2 (6)  9 (28)  1 (3)  Ovary  18  8 (44)  8 (44)  0  1 (6)  8 (44)  4 (22)  14 (78)  9 (50)  Pancreas  40  4 (10)  7 (18)  1 (3)  1 (3)  7 (18)  1 (3)  24 (60)  13 (33)  Prostate  45  2 (4)  1 (2)  4 (9)  0  6 (13)  17 (38)  1 (2)  7 (16)  Stomach  78  12 (15)  20 (26)  1 (1)  2 (3)  22 (28)  4 (5)  52 (67)  32 (41)  Type   No.   CDX2 + CK20, No. (%)  CDX2 + Villin, No. (%)  SATB2 + CK20, No. (%)  SATB2 + Villin, No. (%)  CDX2, No. (%)  SATB2, No. (%)  Villin, No. (%)  CK20, No. (%)  Colorectal  222  205 (92)  213 (96)  199 (90)  206 (93)  220 (99)  211 (95)  215 (97)  207 (93)  Breast  45  0  0  0  0  0  7 (16)  2 (4)  0  Hepatocellular  67  0  0  3 (4)  5 (7)  0  5 (7)  34 (51)  9 (13)  Lung  50  2 (4)  6 (12)  0  3 (6)  6 (12)  6 (12)  29 (58)  8 (16)  Neuroendocrine  32  0  3 (9)  0  1 (3)  9 (28)  2 (6)  9 (28)  1 (3)  Ovary  18  8 (44)  8 (44)  0  1 (6)  8 (44)  4 (22)  14 (78)  9 (50)  Pancreas  40  4 (10)  7 (18)  1 (3)  1 (3)  7 (18)  1 (3)  24 (60)  13 (33)  Prostate  45  2 (4)  1 (2)  4 (9)  0  6 (13)  17 (38)  1 (2)  7 (16)  Stomach  78  12 (15)  20 (26)  1 (1)  2 (3)  22 (28)  4 (5)  52 (67)  32 (41)  CDX2, caudal-type homeobox 2; CK20, cytokeratin 20; CRC, colorectal carcinoma; SATB2, special AT-rich sequence binding protein 2. View Large The sensitivity and specificity for each single marker and each dual stain were determined. SATB2 single stain showed a sensitivity of 95% and a specificity of 88% for CRCs, while CDX2 single stain showed a sensitivity of 99% and a specificity of 85%. CK20 single stain showed a sensitivity of 97% and a specificity of 56% for CRCs, while villin single stain showed a sensitivity of 93% and a specificity of 79%. For dual stains with both stains as positive (SATB2+/CK20+, CDX2+/CK20+, etc), dual stain with SATB2 and CK20 showed a sensitivity of 90% and a specificity of 98% for CRC, while dual stain with CDX2 and CK20 showed a slightly higher sensitivity (92%) but lower specificity (93%). Similarly, dual stain with SATB2 and villin showed a sensitivity of 93% and a specificity of 97% for CRC, while dual stain with CDX2 and villin showed slightly higher sensitivity (96%) but much lower specificity (88%) Table 2 . Table 2 Sensitivity, Specificity, and Overall Percentages of Different Dual Stains and Single Stains (CRC vs Other Tumors) Type   Stain   Sensitivity (95% CI)  Specificity (95% CI)  Positive Predictive Value (95% CI)  Negative Predictive Value (95% CI)  Overall % Correct (95% CI)  Two stains—both  positive  CDX2 + CK20  0.92 (0.88-0.95)  0.93 (0.89-0.95)  0.88 (0.83-0.92)  0.95 (0.93-0.97)  0.92 (0.9-0.94)  CDX2 + villin  0.96 (0.92-0.98)  0.88 (0.84-0.91)  0.83 (0.77-0.87)  0.97 (0.95-0.99)  0.91 (0.88-0.93)  SATB2 + CK20  0.9 (0.85-0.93)  0.98 (0.95-0.99)  0.96 (0.92-0.98)  0.94 (0.91-0.96)  0.95 (0.93-0.96)  SATB2 + villin  0.93 (0.89-0.96)  0.97 (0.94-0.98)  0.94 (0.9-0.97)  0.96 (0.93-0.98)  0.95 (0.93-0.97)  Single stains  CDX2  0.99 (0.97-1)  0.85 (0.8-0.88)  0.79 (0.74-0.84)  0.99 (0.98-1)  0.9 (0.87-0.92)  SATB2  0.95 (0.91-0.98)  0.88 (0.84-0.91)  0.82 (0.77-0.87)  0.97 (0.94-0.98)  0.9 (0.88-0.93)  Villin  0.97 (0.94-0.99)  0.56 (0.51-0.61)  0.57 (0.51-0.62)  0.97 (0.93-0.99)  0.71 (0.67-0.75)  CK20  0.93 (0.89-0.96)  0.79 (0.74-0.83)  0.72 (0.67-0.77)  0.95 (0.92-0.97)  0.84 (0.81-0.87)  Type   Stain   Sensitivity (95% CI)  Specificity (95% CI)  Positive Predictive Value (95% CI)  Negative Predictive Value (95% CI)  Overall % Correct (95% CI)  Two stains—both  positive  CDX2 + CK20  0.92 (0.88-0.95)  0.93 (0.89-0.95)  0.88 (0.83-0.92)  0.95 (0.93-0.97)  0.92 (0.9-0.94)  CDX2 + villin  0.96 (0.92-0.98)  0.88 (0.84-0.91)  0.83 (0.77-0.87)  0.97 (0.95-0.99)  0.91 (0.88-0.93)  SATB2 + CK20  0.9 (0.85-0.93)  0.98 (0.95-0.99)  0.96 (0.92-0.98)  0.94 (0.91-0.96)  0.95 (0.93-0.96)  SATB2 + villin  0.93 (0.89-0.96)  0.97 (0.94-0.98)  0.94 (0.9-0.97)  0.96 (0.93-0.98)  0.95 (0.93-0.97)  Single stains  CDX2  0.99 (0.97-1)  0.85 (0.8-0.88)  0.79 (0.74-0.84)  0.99 (0.98-1)  0.9 (0.87-0.92)  SATB2  0.95 (0.91-0.98)  0.88 (0.84-0.91)  0.82 (0.77-0.87)  0.97 (0.94-0.98)  0.9 (0.88-0.93)  Villin  0.97 (0.94-0.99)  0.56 (0.51-0.61)  0.57 (0.51-0.62)  0.97 (0.93-0.99)  0.71 (0.67-0.75)  CK20  0.93 (0.89-0.96)  0.79 (0.74-0.83)  0.72 (0.67-0.77)  0.95 (0.92-0.97)  0.84 (0.81-0.87)  CDX2, caudal-type homeobox 2; CI, confidence interval; CK20, cytokeratin 20; CRC, colorectal carcinoma; SATB2, special AT-rich sequence binding protein 2. View Large Finally, we directly compared the sensitivity and specificity among different single stains or dual stains. As expected, dual stains showed higher specificity but lower sensitivity than single stains. For single stains, CDX2 showed significantly higher sensitivity than SATB2 but lower specificity, although not statistically significant. For dual stains, there were no strong differences in sensitivity and specificity between SATB2/CK20 and SATB2/villin in our cohort. However, both SATB2/CK20 and SATB2/villin had significantly higher specificities than their counterparts (CDX2/CK20 and CDX2/villin) (SATB2/CK20 vs CDX2/CK20, P = .001; SATB2/villin vs CDX2/villin, P < .001) Table 3 . Table 3 Comparison of Sensitivity and Specificity for All Stains (Sensitivity and Specificity Are Relative to CRC) Stain Comparison   Metric   Difference (95% CI)  P Value   CDX2 + CK20 vs  SATB2 + CK20  Sensitivity  0.03 (–0.03 to 0.09)  .109  Specificity  –0.05 (–0.09 to –0.02)  .001  CDX2 + villin vs  SATB2 + villin  Sensitivity  0.03 (–0.02 to 0.08)  .065  Specificity  –0.09 (–0.13 to –0.05)  <.001  SATB2 + CK20 vs  SATB2 + villin  Sensitivity  –0.03 (–0.09 to 0.03)  .143  Specificity  0.01 (–0.02 to 0.04)  .388  CDX2 vs SATB2  Sensitivity  0.04 (0.01 to 0.08)  .022  Specificity  –0.03 (–0.08 to 0.02)  .241  SATB2 vs SATB2 +  CK20  Sensitivity  0.05 (0 to 0.11)  <.001  Specificity  –0.1 (–0.14 to –0.06)  <.001  SATB2 vs SATB2 +  villin  Sensitivity  0.02 (–0.03 to 0.07)  .063  Specificity  –0.09 (–0.13 to –0.05)  <.001  Stain Comparison   Metric   Difference (95% CI)  P Value   CDX2 + CK20 vs  SATB2 + CK20  Sensitivity  0.03 (–0.03 to 0.09)  .109  Specificity  –0.05 (–0.09 to –0.02)  .001  CDX2 + villin vs  SATB2 + villin  Sensitivity  0.03 (–0.02 to 0.08)  .065  Specificity  –0.09 (–0.13 to –0.05)  <.001  SATB2 + CK20 vs  SATB2 + villin  Sensitivity  –0.03 (–0.09 to 0.03)  .143  Specificity  0.01 (–0.02 to 0.04)  .388  CDX2 vs SATB2  Sensitivity  0.04 (0.01 to 0.08)  .022  Specificity  –0.03 (–0.08 to 0.02)  .241  SATB2 vs SATB2 +  CK20  Sensitivity  0.05 (0 to 0.11)  <.001  Specificity  –0.1 (–0.14 to –0.06)  <.001  SATB2 vs SATB2 +  villin  Sensitivity  0.02 (–0.03 to 0.07)  .063  Specificity  –0.09 (–0.13 to –0.05)  <.001  CDX2, caudal-type homeobox 2; CI, confidence interval; CK20, cytokeratin 20; CRC, colorectal carcinoma; SATB2, special AT-rich sequence binding protein 2. View Large Discussion Many patients with CRC first present with distant metastases, which may require core biopsies to identify tumor origin and direct further clinical management. In such cases, IHC is usually performed to identify/verify the primary site. Pathologists may have very limited materials available for workup, and if multiple IHC stains are necessary, tissue depletion will limit the number of stains. Therefore, it is important to use IHC stains efficiently to determine tumor origin without doing unnecessary stains to waste tissue and increase costs. Two recent studies have demonstrated that SATB2 has a high sensitivity and specificity to determine CRC origin, especially in combination with CK20.12,13 Dual stains or multiplex immunostains are very efficient to use limited materials. In our previous study, we demonstrated that dual stain with one of the nuclear stains SATB2 or CDX2 in combination with the cytoplasmic stain of CK20 or villin was useful in differentiating appendiceal mucinous neoplasms from ovarian mucinous neoplasms.18 In this study, we compared their sensitivities and specificities to determine whether any combination could provide improved diagnostic utility for CRC. Consistent with the previous findings, our data showed SATB2 stained the majority of CRCs while showing very little expression in other tumors with a sensitivity of 95% and a specificity of 88% for CRCs. Furthermore, our study demonstrated that the dual stain with SATB2 and villin (sensitivity, 93%; specificity, 97%) and SATB2 and CK20 (sensitivity, 90%; specificity, 98%) had the highest sensitivities and specificities for CRCs in our study cohort. Finally, we found that dual stain with SATB2/CK20 (or villin) (both positive) had a significantly higher specificity in identifying CRC than dual stain with CDX2/CK20 (villin) (both positive). Our data demonstrate that SATB2 is a useful stain in those cases with a nonclassic immunoprofile (CDX2-negative or CK20-negative immunoprofile). This is very important since this immunoprofile could be seen in upper GI and pancreatobiliary tumors, which can also metastasize to the liver and be confused with metastatic colorectal cancer. The two CDX2-negative CRCs in our series were positive for SATB2. In addition, SATB2 was positive in 12 of 15 CRCs with CDX2-positive and CK20-negative staining. Taken together, these results suggest that SATB2 may be useful to identify CRCs with a nonclassic immunoprofile in a setting of metastasis. The current study is limited by the fact that the sensitivity of the single stains was very high (>90% for all stains), making improvements in sensitivity with dual stains difficult to detect. In addition, we used TMA rather than whole sections to demonstrate the utility of these stains in core needle biopsy specimens. In summary, the current study is one of the few showing the utility of SATB2 in identifying CRC on TMA cores, which mimic small biopsy specimens in the setting of metastatic CRC. In addition, it is the first immunohistochemical analysis with the use of dual stains combining nuclear staining (SATB2 or CDX2) with cytoplasmic staining (CK20 or villin) and directly comparing dual stains of SATB2 and CK20/villin with dual stains of CDX2 and CK20/villin in identifying CRC. Our data suggest that the dual stain with SATB2/CK20 (or villin) shows potential clinical utility in identifying CRC and has higher sensitivity/specificity compared with any single stain or dual stain with CDX2 and CK20 or villin. Dual stains could be helpful in working up specimens with limited tissues. Finally, the use of SATB2 may be useful in those cases with a nonclassic immunoprofile for colorectal adenocarcinoma to help distinguish them from other tumors. References 1. O’Connell JB, Maggard MA, Ko CY. Colon cancer survival rates with the new American Joint Committee on Cancer Sixth Edition staging. J Natl Cancer Inst . 2004; 96: 1420- 1425. Google Scholar CrossRef Search ADS PubMed  2. Hernandez BY, Frierson HF, Moskaluk CAet al.   CK20 and CK7 protein expression in colorectal cancer: demonstration of the utility of a population-based tissue microarray. Hum Pathol . 2005; 36: 275- 281. Google Scholar CrossRef Search ADS PubMed  3. Bayrak R, Yenidünya S, Haltas H. Cytokeratin 7 and cytokeratin 20 expression in colorectal adenocarcinomas. Pathol Res Pract . 2011; 207: 156- 160. Google Scholar CrossRef Search ADS PubMed  4. Tot T. Adenocarcinomas metastatic to the liver: the value of cytokeratins 20 and 7 in the search for unknown primary tumors. Cancer . 1999; 85: 171- 177. Google Scholar CrossRef Search ADS PubMed  5. Kaimaktchiev V, Terracciano L, Tornillo Let al.   The homeobox intestinal differentiation factor CDX2 is selectively expressed in gastrointestinal adenocarcinomas. Mod Pathol . 2004; 17: 1392- 1399. Google Scholar CrossRef Search ADS PubMed  6. Werling RW, Yaziji H, Bacchi CEet al.   CDX2, a highly sensitive and specific marker of adenocarcinomas of intestinal origin: an immunohistochemical survey of 476 primary and metastatic carcinomas. Am J Surg Pathol . 2003; 27: 303- 310. Google Scholar CrossRef Search ADS PubMed  7. Bayrak R, Haltas H, Yenidunya S. The value of CDX2 and cytokeratins 7 and 20 expression in differentiating colorectal adenocarcinomas from extraintestinal gastrointestinal adenocarcinomas: cytokeratin 7–/20+ phenotype is more specific than CDX2 antibody. Diagn Pathol . 2012; 7: 9. Google Scholar CrossRef Search ADS PubMed  8. Saad RS, Ghorab Z, Khalifa MAet al.   CDX2 as a marker for intestinal differentiation: its utility and limitations. World J Gastrointest Surg . 2011; 3: 159- 166. Google Scholar CrossRef Search ADS PubMed  9. Tot T. Identifying colorectal metastases in liver biopsies: the novel CDX2 antibody is less specific than the cytokeratin 20+/7– phenotype. Med Sci Monit . 2004; 10: BR139- BR143. Google Scholar PubMed  10. Björling E, Lindskog C, Oksvold Pet al.   A web-based tool for in silico biomarker discovery based on tissue-specific protein profiles in normal and cancer tissues. Mol Cell Proteomics . 2008; 7: 825- 844. Google Scholar CrossRef Search ADS PubMed  11. FitzPatrick DR, Carr IM, McLaren Let al.   Identification of SATB2 as the cleft palate gene on 2q32-q33. Hum Mol Genet . 2003; 12: 2491- 2501. Google Scholar CrossRef Search ADS PubMed  12. Magnusson K, de Wit M, Brennan DJet al.   SATB2 in combination with cytokeratin 20 identifies over 95% of all colorectal carcinomas. Am J Surg Pathol . 2011; 35: 937- 948. Google Scholar CrossRef Search ADS PubMed  13. Dragomir A, de Wit M, Johansson Cet al.   The role of SATB2 as a diagnostic marker for tumors of colorectal origin: results of a pathology-based clinical prospective study. Am J Clin Pathol . 2014; 141: 630- 638. Google Scholar CrossRef Search ADS PubMed  14. Wang S, Zhou J, Wang XYet al.   Down-regulated expression of SATB2 is associated with metastasis and poor prognosis in colorectal cancer. J Pathol . 2009; 219: 114- 122. Google Scholar CrossRef Search ADS PubMed  15. Eberhard J, Gaber A, Wangefjord Set al.   A cohort study of the prognostic and treatment predictive value of SATB2 expression in colorectal cancer. Br J Cancer . 2012; 106: 931- 938. Google Scholar CrossRef Search ADS PubMed  16. Conner JR, Hornick JL. SATB2 is a novel marker of osteoblastic differentiation in bone and soft tissue tumours. Histopathology . 2013; 63: 36- 49. Google Scholar CrossRef Search ADS PubMed  17. Righi A, Gambarotti M, Longo Set al.   Small cell osteosarcoma: clinicopathologic, immunohistochemical, and molecular analysis of 36 cases. Am J Surg Pathol . 2015; 39: 691- 699. Google Scholar CrossRef Search ADS PubMed  18. Li Z, Roth R, Rock JBet al.   Dual immunostain with SATB2 and CK20 differentiates appendiceal mucinous neoplasms from ovarian mucinous neoplasms. Am J Clin Pathol . 2017; 147: 484- 491. Google Scholar CrossRef Search ADS PubMed  19. Newcombe RG. Improved confidence intervals for the difference between binomial proportions based on paired data. Stat Med . 1998; 17: 2635- 2650. Google Scholar CrossRef Search ADS PubMed  20. Newcombe RG. Simultaneous comparison of sensitivity and specificity of two tests in the paired design: a straightforward graphical approach. Stat Med . 2001; 20: 907- 915. Google Scholar CrossRef Search ADS PubMed  © American Society for Clinical Pathology, 2018. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png American Journal of Clinical Pathology Oxford University Press

Dual Stain With SATB2 and CK20/Villin Is Useful to Distinguish Colorectal Carcinomas From Other Tumors

Loading next page...
 
/lp/ou_press/dual-stain-with-satb2-and-ck20-villin-is-useful-to-distinguish-HZ8uHP2R00
Publisher
Oxford University Press
Copyright
© American Society for Clinical Pathology, 2018. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
0002-9173
eISSN
1943-7722
D.O.I.
10.1093/ajcp/aqx160
Publisher site
See Article on Publisher Site

Abstract

Abstract Objectives Small sample size limits the number of immunostains that may be attempted in colorectal carcinoma (CRC) biopsy specimens. We investigated the utility of dual stain with special AT-rich sequence binding protein 2 (SATB2) or caudal-type homeobox 2 (CDX2) and cytokeratin 20 (CK20) or villin in identifying CRC. Methods Tissue microarrays with 222 CRCs and 375 other carcinomas were built. Dual stain was performed pairing nuclear stains CDX2 or SATB2 with CK20 or villin. Results All four single stains showed excellent sensitivity (93%-99%) but variable specificity (56%-88%) for CRC. All four dual stains also showed excellent sensitivity (90%-96%) while much improved specificity (88%-98%) compared with single stains. SATB2 dual stain (with CK20 or villin) showed a higher specificity than CDX2 dual stain (with CK20 or villin) with a comparable sensitivity. Conclusions SATB2 dual stain shows the greatest potential clinical utility in identifying CRC and is superior to CDX2 dual stain. More important, SATB2 dual stain could be helpful for specimens with limited tissues or those having a nonclassic staining pattern. Dual immunostain, SATB2, CDX2, CK20, Villin, Colorectal tumor Colorectal carcinoma (CRC) is one of the most common cancers worldwide and in the United States.1 A significant proportion of patients with CRC initially present with distant metastases, frequently requiring core biopsy for diagnosis and to direct further management. In such case, immunohistochemistry (IHC) is usually performed to identify the primary site. Cytokeratin 20 (CK20) is an intermediate filament protein expressed in gastrointestinal (GI) tract glandular epithelial cells and is routinely used as a diagnostic marker for CRC.2-4 CK20 is a highly sensitive marker for CRC. However, its specificity is low due to its expression in several other types of adenocarcinomas.2-4 Similar results have also been found with villin, another marker that has been used in the GI tract. Caudal-type homeobox 2 (CDX2) is a transcription factor regulating intestinal epithelial cell differentiation that is strongly expressed in up to 100% of CRCs but also expressed in gastric carcinomas (18%) and ovarian mucinous tumors (20%).5-10 Special AT-rich sequence binding protein 2 (SATB2) is a DNA-binding protein expressed in glandular cells lining the lower GI tract.11 SATB2 is a highly sensitive and also specific marker for CRC, and SATB2 in combination with CK20 can identify 97% of CRCs.12 The same study also revealed that only 3.3% of ovarian carcinomas and 5.7% of lung adenocarcinomas were positive for SATB2, while all gastric and pancreatic carcinomas were all negative for SATB2.12 Additional study has demonstrated that SATB2 showed a high sensitivity (93%) and specificity (77%) to identify a metastasis as having a colorectal origin, and in combination with CK7 and CK20, the specificity increased to 100%.13 Decreased SATB2 expression level has been shown to associate with metastasis and poor prognosis in CRCs.14,15 SATB2 is also a marker of osteoblastic differentiation in benign and malignant mesenchymal tumors.16,17 Metastatic colorectal carcinoma biopsy specimens sometimes have limited materials for further workup; therefore, dual stains or multiplex immunostains are very efficient to use limited materials. Given the high sensitivity and specificity of SATB2 in CRC and the challenge in identifying the primary site for metastatic CRC, we were interested in its utility in a dual stain with either CK20 or villin compared with CDX2 in a dual stain with the same two markers in identifying CRC on limited materials. In this study, we explored dual stains with one of the nuclear stains SATB2 or CDX2 in combination with the cytoplasmic stain of CK20 or villin and compared their sensitivities and specificities to determine whether any combination could provide improved diagnostic utility for CRC. Materials and Methods Case Selection and Microarray Construction After institutional review board approval at Ohio State University (OSU), an electronic pathology archive database search was performed to select CRCs diagnosed from 2000 to 2011 at the OSU Wexner Medical Center. In total, 222 CRCs and 375 other carcinomas (45 breast, 67 hepatocellular, 50 lung, 32 neuroendocrine including 31 from pancreas and one from small bowel, 18 ovarian mucinous tumors, 40 pancreas, 45 prostate, and 78 gastric) were identified, and the diagnosis was confirmed after reviewing pathologic slides and correlating with patients’ clinical history. All cases in our cohort were from the primary site of origin. A tumor block representative of the primary tumor was collected from each case, and tissue microarrays (TMAs) were constructed with duplicate cores of 1 mm for each tumor. Dual Immunostains and Scoring dual stains were performed on TMAs by pairing nuclear stains CDX2 (1:300; BioGenex, San Ramon, CA) or SATB2 (1:50; Santa Cruz, Santa Cruz, CA) with cytoplasmic stains CK20 (1:200; Dako, Carpenteria, CA) or villin (1:50; Cell Marque, Rocklin, CA). The detailed staining protocols were previously described.18 Briefly, 4-μm sections were cut from TMA blocks. CDX2 and SATB2 were labeled as brown color and CK20 and villin were labeled as red color. A Bond Polymer Refine Detection Kit (Leica Biosystems, Buffalo Grove, IL) was used as the brown chromogen, and Bond Polymer Refine Red Detection Kit (Leica Biosystems) was used as the red chromogen. Stained slides were independently reviewed by three pathologists (J.B.R., R.R., and W.L.F.), and a consensus staining result was reached for each case. Staining results were scored as positive or negative with a positive threshold of staining of 5% or more of tumor cells. Representative images are demonstrated in Image 1 and Image 2 . Image 1 View largeDownload slide Representative images of dual stains with SATB2 or CDX2 and CK20 or villin in moderately differentiated and poorly differentiated colorectal carcinomas. SATB2 and CDX2 showed brown nuclear staining, while CK20 and villin showed red cytoplasmic staining (×200). CDX2, caudal-type homeobox 2; CK20, cytokeratin 20; SATB2, special AT-rich sequence binding protein 2. Image 1 View largeDownload slide Representative images of dual stains with SATB2 or CDX2 and CK20 or villin in moderately differentiated and poorly differentiated colorectal carcinomas. SATB2 and CDX2 showed brown nuclear staining, while CK20 and villin showed red cytoplasmic staining (×200). CDX2, caudal-type homeobox 2; CK20, cytokeratin 20; SATB2, special AT-rich sequence binding protein 2. Image 2 View largeDownload slide Representative images of dual stains with CDX2/SATB2 and CK20/villin in different carcinomas. A colonic adenocarcinoma showed negative CDX2 and CK20 staining with CDX2/CK20 dual stain but positive SATB2 staining with SATB2/CK20 dual stain. A lung adenocarcinoma showed positive CDX2 staining with CDX2/CK20 dual stain but negative SATB2 and CK20 staining with SATB2/CK20 dual stain. A pancreatic adenocarcinoma showed positive CDX2 staining with CDX2/villin dual stain but negative SATB2 and villin staining with SATB2/villin dual stain. A stomach adenocarcinoma showed positive CDX2 and CK20 staining with CDX2/CK20 dual stain but negative SATB2 and positive CK20 staining with SATB2/CK20 dual stain (×200). CDX2, caudal-type homeobox 2; CK20, cytokeratin 20; SATB2, special AT-rich sequence binding protein 2. Image 2 View largeDownload slide Representative images of dual stains with CDX2/SATB2 and CK20/villin in different carcinomas. A colonic adenocarcinoma showed negative CDX2 and CK20 staining with CDX2/CK20 dual stain but positive SATB2 staining with SATB2/CK20 dual stain. A lung adenocarcinoma showed positive CDX2 staining with CDX2/CK20 dual stain but negative SATB2 and CK20 staining with SATB2/CK20 dual stain. A pancreatic adenocarcinoma showed positive CDX2 staining with CDX2/villin dual stain but negative SATB2 and villin staining with SATB2/villin dual stain. A stomach adenocarcinoma showed positive CDX2 and CK20 staining with CDX2/CK20 dual stain but negative SATB2 and positive CK20 staining with SATB2/CK20 dual stain (×200). CDX2, caudal-type homeobox 2; CK20, cytokeratin 20; SATB2, special AT-rich sequence binding protein 2. Statistical Analysis The sensitivity, specificity, and overall percentage correct in identifying CRCs were calculated with exact binomial 95% confidence intervals (CIs) for each stain and dual stain separately. Sensitivity and specificity were directly compared between selected stains using the McNemar test for paired comparisons; exact P values were calculated. The 95% CIs for the differences between sensitivities or specificities were calculated using methods based on the score interval as previously described.19,20 All analyses were performed using SAS/STAT software version 9.4 of the SAS System for Windows (SAS Institute, Cary, NC). Results We first evaluated the staining results in 222 CRCs and 375 other tumors for each marker individually. As shown in Table 1 , SATB2 nuclear staining was present in 95% of CRCs, followed by 38% of prostate cancers, 22% of ovarian cancers, 16% of breast cancers, 12% of lung cancers, 7% of hepatocellular carcinomas, 6% of neuroendocrine tumors, 5% of gastric cancers, and 3% of pancreatic cancers. CK20 and villin cytoplasmic staining was present in 93% and 97% of CRCs, respectively, and CDX2 nuclear staining was seen in 99% of CRCs. In our cohort, 197 CRCs were positive for all of the following three stains: CDX2, CK20, and SATB2. Two CRCs (0.9%) were CDX2 negative, and both were positive for SATB2 and CK20. Three (1.4%) CRCs were CK20 negative and STAB2 negative but CDX2 positive. Twelve (5.4%) CRCs were CDX2 positive and SATB2 positive but CK20 negative. Eight (3.6%) CRCs were SATB2 negative but CDX2 positive and CK20 positive. There was no case that was both CDX2 negative and CK20 negative in our cohort. Next we evaluated the dual staining results. Dual stain with SATB2 and CK20 showed positivity for both markers in 90% of CRCs and in 0% to 9% of other tumors. Dual stain with SATB2 and villin showed positivity for both markers in 93% of CRCs and 0% to 7% of other tumors. Although dual stains with CDX2 and CK20/villin showed positivity for both markers in up to 96% of CRCs, they were positive in up to 44% of other tumors (Table 1). Table 1 Frequency of Single Stains and Dual Stains in Different Primary Tumors Type   No.   CDX2 + CK20, No. (%)  CDX2 + Villin, No. (%)  SATB2 + CK20, No. (%)  SATB2 + Villin, No. (%)  CDX2, No. (%)  SATB2, No. (%)  Villin, No. (%)  CK20, No. (%)  Colorectal  222  205 (92)  213 (96)  199 (90)  206 (93)  220 (99)  211 (95)  215 (97)  207 (93)  Breast  45  0  0  0  0  0  7 (16)  2 (4)  0  Hepatocellular  67  0  0  3 (4)  5 (7)  0  5 (7)  34 (51)  9 (13)  Lung  50  2 (4)  6 (12)  0  3 (6)  6 (12)  6 (12)  29 (58)  8 (16)  Neuroendocrine  32  0  3 (9)  0  1 (3)  9 (28)  2 (6)  9 (28)  1 (3)  Ovary  18  8 (44)  8 (44)  0  1 (6)  8 (44)  4 (22)  14 (78)  9 (50)  Pancreas  40  4 (10)  7 (18)  1 (3)  1 (3)  7 (18)  1 (3)  24 (60)  13 (33)  Prostate  45  2 (4)  1 (2)  4 (9)  0  6 (13)  17 (38)  1 (2)  7 (16)  Stomach  78  12 (15)  20 (26)  1 (1)  2 (3)  22 (28)  4 (5)  52 (67)  32 (41)  Type   No.   CDX2 + CK20, No. (%)  CDX2 + Villin, No. (%)  SATB2 + CK20, No. (%)  SATB2 + Villin, No. (%)  CDX2, No. (%)  SATB2, No. (%)  Villin, No. (%)  CK20, No. (%)  Colorectal  222  205 (92)  213 (96)  199 (90)  206 (93)  220 (99)  211 (95)  215 (97)  207 (93)  Breast  45  0  0  0  0  0  7 (16)  2 (4)  0  Hepatocellular  67  0  0  3 (4)  5 (7)  0  5 (7)  34 (51)  9 (13)  Lung  50  2 (4)  6 (12)  0  3 (6)  6 (12)  6 (12)  29 (58)  8 (16)  Neuroendocrine  32  0  3 (9)  0  1 (3)  9 (28)  2 (6)  9 (28)  1 (3)  Ovary  18  8 (44)  8 (44)  0  1 (6)  8 (44)  4 (22)  14 (78)  9 (50)  Pancreas  40  4 (10)  7 (18)  1 (3)  1 (3)  7 (18)  1 (3)  24 (60)  13 (33)  Prostate  45  2 (4)  1 (2)  4 (9)  0  6 (13)  17 (38)  1 (2)  7 (16)  Stomach  78  12 (15)  20 (26)  1 (1)  2 (3)  22 (28)  4 (5)  52 (67)  32 (41)  CDX2, caudal-type homeobox 2; CK20, cytokeratin 20; CRC, colorectal carcinoma; SATB2, special AT-rich sequence binding protein 2. View Large The sensitivity and specificity for each single marker and each dual stain were determined. SATB2 single stain showed a sensitivity of 95% and a specificity of 88% for CRCs, while CDX2 single stain showed a sensitivity of 99% and a specificity of 85%. CK20 single stain showed a sensitivity of 97% and a specificity of 56% for CRCs, while villin single stain showed a sensitivity of 93% and a specificity of 79%. For dual stains with both stains as positive (SATB2+/CK20+, CDX2+/CK20+, etc), dual stain with SATB2 and CK20 showed a sensitivity of 90% and a specificity of 98% for CRC, while dual stain with CDX2 and CK20 showed a slightly higher sensitivity (92%) but lower specificity (93%). Similarly, dual stain with SATB2 and villin showed a sensitivity of 93% and a specificity of 97% for CRC, while dual stain with CDX2 and villin showed slightly higher sensitivity (96%) but much lower specificity (88%) Table 2 . Table 2 Sensitivity, Specificity, and Overall Percentages of Different Dual Stains and Single Stains (CRC vs Other Tumors) Type   Stain   Sensitivity (95% CI)  Specificity (95% CI)  Positive Predictive Value (95% CI)  Negative Predictive Value (95% CI)  Overall % Correct (95% CI)  Two stains—both  positive  CDX2 + CK20  0.92 (0.88-0.95)  0.93 (0.89-0.95)  0.88 (0.83-0.92)  0.95 (0.93-0.97)  0.92 (0.9-0.94)  CDX2 + villin  0.96 (0.92-0.98)  0.88 (0.84-0.91)  0.83 (0.77-0.87)  0.97 (0.95-0.99)  0.91 (0.88-0.93)  SATB2 + CK20  0.9 (0.85-0.93)  0.98 (0.95-0.99)  0.96 (0.92-0.98)  0.94 (0.91-0.96)  0.95 (0.93-0.96)  SATB2 + villin  0.93 (0.89-0.96)  0.97 (0.94-0.98)  0.94 (0.9-0.97)  0.96 (0.93-0.98)  0.95 (0.93-0.97)  Single stains  CDX2  0.99 (0.97-1)  0.85 (0.8-0.88)  0.79 (0.74-0.84)  0.99 (0.98-1)  0.9 (0.87-0.92)  SATB2  0.95 (0.91-0.98)  0.88 (0.84-0.91)  0.82 (0.77-0.87)  0.97 (0.94-0.98)  0.9 (0.88-0.93)  Villin  0.97 (0.94-0.99)  0.56 (0.51-0.61)  0.57 (0.51-0.62)  0.97 (0.93-0.99)  0.71 (0.67-0.75)  CK20  0.93 (0.89-0.96)  0.79 (0.74-0.83)  0.72 (0.67-0.77)  0.95 (0.92-0.97)  0.84 (0.81-0.87)  Type   Stain   Sensitivity (95% CI)  Specificity (95% CI)  Positive Predictive Value (95% CI)  Negative Predictive Value (95% CI)  Overall % Correct (95% CI)  Two stains—both  positive  CDX2 + CK20  0.92 (0.88-0.95)  0.93 (0.89-0.95)  0.88 (0.83-0.92)  0.95 (0.93-0.97)  0.92 (0.9-0.94)  CDX2 + villin  0.96 (0.92-0.98)  0.88 (0.84-0.91)  0.83 (0.77-0.87)  0.97 (0.95-0.99)  0.91 (0.88-0.93)  SATB2 + CK20  0.9 (0.85-0.93)  0.98 (0.95-0.99)  0.96 (0.92-0.98)  0.94 (0.91-0.96)  0.95 (0.93-0.96)  SATB2 + villin  0.93 (0.89-0.96)  0.97 (0.94-0.98)  0.94 (0.9-0.97)  0.96 (0.93-0.98)  0.95 (0.93-0.97)  Single stains  CDX2  0.99 (0.97-1)  0.85 (0.8-0.88)  0.79 (0.74-0.84)  0.99 (0.98-1)  0.9 (0.87-0.92)  SATB2  0.95 (0.91-0.98)  0.88 (0.84-0.91)  0.82 (0.77-0.87)  0.97 (0.94-0.98)  0.9 (0.88-0.93)  Villin  0.97 (0.94-0.99)  0.56 (0.51-0.61)  0.57 (0.51-0.62)  0.97 (0.93-0.99)  0.71 (0.67-0.75)  CK20  0.93 (0.89-0.96)  0.79 (0.74-0.83)  0.72 (0.67-0.77)  0.95 (0.92-0.97)  0.84 (0.81-0.87)  CDX2, caudal-type homeobox 2; CI, confidence interval; CK20, cytokeratin 20; CRC, colorectal carcinoma; SATB2, special AT-rich sequence binding protein 2. View Large Finally, we directly compared the sensitivity and specificity among different single stains or dual stains. As expected, dual stains showed higher specificity but lower sensitivity than single stains. For single stains, CDX2 showed significantly higher sensitivity than SATB2 but lower specificity, although not statistically significant. For dual stains, there were no strong differences in sensitivity and specificity between SATB2/CK20 and SATB2/villin in our cohort. However, both SATB2/CK20 and SATB2/villin had significantly higher specificities than their counterparts (CDX2/CK20 and CDX2/villin) (SATB2/CK20 vs CDX2/CK20, P = .001; SATB2/villin vs CDX2/villin, P < .001) Table 3 . Table 3 Comparison of Sensitivity and Specificity for All Stains (Sensitivity and Specificity Are Relative to CRC) Stain Comparison   Metric   Difference (95% CI)  P Value   CDX2 + CK20 vs  SATB2 + CK20  Sensitivity  0.03 (–0.03 to 0.09)  .109  Specificity  –0.05 (–0.09 to –0.02)  .001  CDX2 + villin vs  SATB2 + villin  Sensitivity  0.03 (–0.02 to 0.08)  .065  Specificity  –0.09 (–0.13 to –0.05)  <.001  SATB2 + CK20 vs  SATB2 + villin  Sensitivity  –0.03 (–0.09 to 0.03)  .143  Specificity  0.01 (–0.02 to 0.04)  .388  CDX2 vs SATB2  Sensitivity  0.04 (0.01 to 0.08)  .022  Specificity  –0.03 (–0.08 to 0.02)  .241  SATB2 vs SATB2 +  CK20  Sensitivity  0.05 (0 to 0.11)  <.001  Specificity  –0.1 (–0.14 to –0.06)  <.001  SATB2 vs SATB2 +  villin  Sensitivity  0.02 (–0.03 to 0.07)  .063  Specificity  –0.09 (–0.13 to –0.05)  <.001  Stain Comparison   Metric   Difference (95% CI)  P Value   CDX2 + CK20 vs  SATB2 + CK20  Sensitivity  0.03 (–0.03 to 0.09)  .109  Specificity  –0.05 (–0.09 to –0.02)  .001  CDX2 + villin vs  SATB2 + villin  Sensitivity  0.03 (–0.02 to 0.08)  .065  Specificity  –0.09 (–0.13 to –0.05)  <.001  SATB2 + CK20 vs  SATB2 + villin  Sensitivity  –0.03 (–0.09 to 0.03)  .143  Specificity  0.01 (–0.02 to 0.04)  .388  CDX2 vs SATB2  Sensitivity  0.04 (0.01 to 0.08)  .022  Specificity  –0.03 (–0.08 to 0.02)  .241  SATB2 vs SATB2 +  CK20  Sensitivity  0.05 (0 to 0.11)  <.001  Specificity  –0.1 (–0.14 to –0.06)  <.001  SATB2 vs SATB2 +  villin  Sensitivity  0.02 (–0.03 to 0.07)  .063  Specificity  –0.09 (–0.13 to –0.05)  <.001  CDX2, caudal-type homeobox 2; CI, confidence interval; CK20, cytokeratin 20; CRC, colorectal carcinoma; SATB2, special AT-rich sequence binding protein 2. View Large Discussion Many patients with CRC first present with distant metastases, which may require core biopsies to identify tumor origin and direct further clinical management. In such cases, IHC is usually performed to identify/verify the primary site. Pathologists may have very limited materials available for workup, and if multiple IHC stains are necessary, tissue depletion will limit the number of stains. Therefore, it is important to use IHC stains efficiently to determine tumor origin without doing unnecessary stains to waste tissue and increase costs. Two recent studies have demonstrated that SATB2 has a high sensitivity and specificity to determine CRC origin, especially in combination with CK20.12,13 Dual stains or multiplex immunostains are very efficient to use limited materials. In our previous study, we demonstrated that dual stain with one of the nuclear stains SATB2 or CDX2 in combination with the cytoplasmic stain of CK20 or villin was useful in differentiating appendiceal mucinous neoplasms from ovarian mucinous neoplasms.18 In this study, we compared their sensitivities and specificities to determine whether any combination could provide improved diagnostic utility for CRC. Consistent with the previous findings, our data showed SATB2 stained the majority of CRCs while showing very little expression in other tumors with a sensitivity of 95% and a specificity of 88% for CRCs. Furthermore, our study demonstrated that the dual stain with SATB2 and villin (sensitivity, 93%; specificity, 97%) and SATB2 and CK20 (sensitivity, 90%; specificity, 98%) had the highest sensitivities and specificities for CRCs in our study cohort. Finally, we found that dual stain with SATB2/CK20 (or villin) (both positive) had a significantly higher specificity in identifying CRC than dual stain with CDX2/CK20 (villin) (both positive). Our data demonstrate that SATB2 is a useful stain in those cases with a nonclassic immunoprofile (CDX2-negative or CK20-negative immunoprofile). This is very important since this immunoprofile could be seen in upper GI and pancreatobiliary tumors, which can also metastasize to the liver and be confused with metastatic colorectal cancer. The two CDX2-negative CRCs in our series were positive for SATB2. In addition, SATB2 was positive in 12 of 15 CRCs with CDX2-positive and CK20-negative staining. Taken together, these results suggest that SATB2 may be useful to identify CRCs with a nonclassic immunoprofile in a setting of metastasis. The current study is limited by the fact that the sensitivity of the single stains was very high (>90% for all stains), making improvements in sensitivity with dual stains difficult to detect. In addition, we used TMA rather than whole sections to demonstrate the utility of these stains in core needle biopsy specimens. In summary, the current study is one of the few showing the utility of SATB2 in identifying CRC on TMA cores, which mimic small biopsy specimens in the setting of metastatic CRC. In addition, it is the first immunohistochemical analysis with the use of dual stains combining nuclear staining (SATB2 or CDX2) with cytoplasmic staining (CK20 or villin) and directly comparing dual stains of SATB2 and CK20/villin with dual stains of CDX2 and CK20/villin in identifying CRC. Our data suggest that the dual stain with SATB2/CK20 (or villin) shows potential clinical utility in identifying CRC and has higher sensitivity/specificity compared with any single stain or dual stain with CDX2 and CK20 or villin. Dual stains could be helpful in working up specimens with limited tissues. Finally, the use of SATB2 may be useful in those cases with a nonclassic immunoprofile for colorectal adenocarcinoma to help distinguish them from other tumors. References 1. O’Connell JB, Maggard MA, Ko CY. Colon cancer survival rates with the new American Joint Committee on Cancer Sixth Edition staging. J Natl Cancer Inst . 2004; 96: 1420- 1425. Google Scholar CrossRef Search ADS PubMed  2. Hernandez BY, Frierson HF, Moskaluk CAet al.   CK20 and CK7 protein expression in colorectal cancer: demonstration of the utility of a population-based tissue microarray. Hum Pathol . 2005; 36: 275- 281. Google Scholar CrossRef Search ADS PubMed  3. Bayrak R, Yenidünya S, Haltas H. Cytokeratin 7 and cytokeratin 20 expression in colorectal adenocarcinomas. Pathol Res Pract . 2011; 207: 156- 160. Google Scholar CrossRef Search ADS PubMed  4. Tot T. Adenocarcinomas metastatic to the liver: the value of cytokeratins 20 and 7 in the search for unknown primary tumors. Cancer . 1999; 85: 171- 177. Google Scholar CrossRef Search ADS PubMed  5. Kaimaktchiev V, Terracciano L, Tornillo Let al.   The homeobox intestinal differentiation factor CDX2 is selectively expressed in gastrointestinal adenocarcinomas. Mod Pathol . 2004; 17: 1392- 1399. Google Scholar CrossRef Search ADS PubMed  6. Werling RW, Yaziji H, Bacchi CEet al.   CDX2, a highly sensitive and specific marker of adenocarcinomas of intestinal origin: an immunohistochemical survey of 476 primary and metastatic carcinomas. Am J Surg Pathol . 2003; 27: 303- 310. Google Scholar CrossRef Search ADS PubMed  7. Bayrak R, Haltas H, Yenidunya S. The value of CDX2 and cytokeratins 7 and 20 expression in differentiating colorectal adenocarcinomas from extraintestinal gastrointestinal adenocarcinomas: cytokeratin 7–/20+ phenotype is more specific than CDX2 antibody. Diagn Pathol . 2012; 7: 9. Google Scholar CrossRef Search ADS PubMed  8. Saad RS, Ghorab Z, Khalifa MAet al.   CDX2 as a marker for intestinal differentiation: its utility and limitations. World J Gastrointest Surg . 2011; 3: 159- 166. Google Scholar CrossRef Search ADS PubMed  9. Tot T. Identifying colorectal metastases in liver biopsies: the novel CDX2 antibody is less specific than the cytokeratin 20+/7– phenotype. Med Sci Monit . 2004; 10: BR139- BR143. Google Scholar PubMed  10. Björling E, Lindskog C, Oksvold Pet al.   A web-based tool for in silico biomarker discovery based on tissue-specific protein profiles in normal and cancer tissues. Mol Cell Proteomics . 2008; 7: 825- 844. Google Scholar CrossRef Search ADS PubMed  11. FitzPatrick DR, Carr IM, McLaren Let al.   Identification of SATB2 as the cleft palate gene on 2q32-q33. Hum Mol Genet . 2003; 12: 2491- 2501. Google Scholar CrossRef Search ADS PubMed  12. Magnusson K, de Wit M, Brennan DJet al.   SATB2 in combination with cytokeratin 20 identifies over 95% of all colorectal carcinomas. Am J Surg Pathol . 2011; 35: 937- 948. Google Scholar CrossRef Search ADS PubMed  13. Dragomir A, de Wit M, Johansson Cet al.   The role of SATB2 as a diagnostic marker for tumors of colorectal origin: results of a pathology-based clinical prospective study. Am J Clin Pathol . 2014; 141: 630- 638. Google Scholar CrossRef Search ADS PubMed  14. Wang S, Zhou J, Wang XYet al.   Down-regulated expression of SATB2 is associated with metastasis and poor prognosis in colorectal cancer. J Pathol . 2009; 219: 114- 122. Google Scholar CrossRef Search ADS PubMed  15. Eberhard J, Gaber A, Wangefjord Set al.   A cohort study of the prognostic and treatment predictive value of SATB2 expression in colorectal cancer. Br J Cancer . 2012; 106: 931- 938. Google Scholar CrossRef Search ADS PubMed  16. Conner JR, Hornick JL. SATB2 is a novel marker of osteoblastic differentiation in bone and soft tissue tumours. Histopathology . 2013; 63: 36- 49. Google Scholar CrossRef Search ADS PubMed  17. Righi A, Gambarotti M, Longo Set al.   Small cell osteosarcoma: clinicopathologic, immunohistochemical, and molecular analysis of 36 cases. Am J Surg Pathol . 2015; 39: 691- 699. Google Scholar CrossRef Search ADS PubMed  18. Li Z, Roth R, Rock JBet al.   Dual immunostain with SATB2 and CK20 differentiates appendiceal mucinous neoplasms from ovarian mucinous neoplasms. Am J Clin Pathol . 2017; 147: 484- 491. Google Scholar CrossRef Search ADS PubMed  19. Newcombe RG. Improved confidence intervals for the difference between binomial proportions based on paired data. Stat Med . 1998; 17: 2635- 2650. Google Scholar CrossRef Search ADS PubMed  20. Newcombe RG. Simultaneous comparison of sensitivity and specificity of two tests in the paired design: a straightforward graphical approach. Stat Med . 2001; 20: 907- 915. Google Scholar CrossRef Search ADS PubMed  © American Society for Clinical Pathology, 2018. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

American Journal of Clinical PathologyOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off