An unusual case of desmoplastic melanoma containing an osteoclast-like giant cell-rich nodule.

The authors describe a case of a 5 cm mixed desmoplastic melanoma occurring on the cheek of an 88-year-old white woman. The epidermis showed the features of lentigo maligna. Within the dermis, there was a mixed desmoplastic melanoma with 2 components. The first component consisted of infiltrative malignant spindled cells with prominent stromal fibrosis and had the typical appearance of desmoplastic melanoma. The second component was within the deep half of the tumor and consisted of a densely cellular nodule composed of spindled melanocytes admixed with many osteoclast-like giant cells. There was a peripheral neurotropism and tumor invaded bone. The Breslow thickness was 14 mm. On followup, a sacral metastasis was discovered, which had a similar morphology to the deep cellular nodule. Immunohistochemistry of spindled cells both inside and outside the nodule showed S100 positivity with the absence of other melanocytic markers (HMB-45, Melan-A). Smooth muscle actin and p63 were focally positive. The osteoclast-like giant cells expressed CD68 and MiTF. Array comparative genomic hybridization of the typical desmoplastic melanoma region had a flat profile, whereas the cellular osteoclast-like giant cell­rich region displayed important cytogenetic anomalies, some of which have been previously described in melanomas. The main array comparative genomic hybridization findings were confirmed by fluorescence in situ hybridization using specific probes. The differences in morphology and molecular cytogenetics between the 2 areas suggest that these might represent the progression or emergence of a more aggressive clone within the tumor. Subsequent metastatic spread to the bone may be a result of accumulated cytogenetic abnormalities.

ALK and ROS1 overexpression is very rare in colorectal adenocarcinoma.

Crizotinib, a small molecule tyrosine kinase inhibitor, has shown tremendous promise in the treatment of lung adenocarcinomas harboring either ALK or ROS1 rearrangements. Recently, small studies of colorectal carcinomas (CRCs) have suggested an incidence of EML4-ALK translocations of 0.4% to 2.4% and FIG-ROS1 translocations of 0.8%. In lung cancer, screening immunohistochemical staining for ALK and ROS1 has been validated as highly sensitive for these translocations, but this has not been investigated in CRC. We therefore sought to investigate the incidence of ALK and ROS1 overexpression as detected by immunohistochemical staining in a large cohort of CRCs. Of the 1889 CRCs, only 1 case (0.05%) demonstrated diffuse strong positive staining for ALK, whereas 14 (0.7%) showed weak nonspecific staining; the remainder were negative. The 1 positive case was confirmed to harbor an ALK rearrangement by fluorescent in situ hybridization (FISH), whereas the 14 tumors with weak staining were FISH-negative. The ALK positive case demonstrated positive expression in all dysplastic and malignant cells indicating that the translocation was an early clonal event. No cases were positive for ROS1 by immunohistochemical staining, although 2 cases did show some nonspecific staining and were shown to be negative for ROS1 translocation by FISH. We conclude that although diffuse strong positive staining for ALK is likely to be highly specific for ALK rearrangement in CRC, both ALK and ROS1 immunohistochemical staining are very low-yield tests and difficult to justify in the routine clinical setting.

EGFR mutation specific immunohistochemistry is a useful adjunct which helps to identify false negative mutation testing in lung cancer.

Mutations in EGFR guide treatment in non-small cell lung cancer (NSCLC). The most common mutations, exon 19 (delE746-A750) and exon 21 (L858R), can be identified by mutation specific immunohistochemistry (IHC). We present our prospective experience of universal reflex IHC and molecular testing in non-squamous NSCLC in the routine clinical setting.A total of 411 specimens from 332 patients were encountered over two years. Of these, 326 (98%) patients underwent EGFR IHC, 15 (5%) were positive for exon 19 deletions and 27 (8%) for exon 21 (L858R); 244 (73%) patients underwent molecular testing. Seventy-six mutations in 64 patients (19% of all patients encountered; 26% with sufficient material for testing) were identified. These comprised nine exon 18 (G719X) mutations, three also with exon 20 mutations; 24 exon 19 deletions, six also with exon 20 mutations; 23 exon 21 (L858R), three also with exon 20 mutations; and 8 exon 20 alone.All 15 exon 19 IHC positive patients were proven mutated (100% specificity, 63% sensitivity). Twenty-two of 27 exon 21 IHC positive cases were proven mutated while three patients had insufficient material for molecular testing (92% specificity, 96% sensitivity). The overall specificity and sensitivity of IHC for any EGFR mutation was 95% and 58%. Five patients initially thought to be wild type for EGFR but IHC positive underwent repeat molecular testing because of the discrepancy which confirmed the IHC result in three cases (60%).We conclude IHC is very specific but not sensitive. Whilst IHC cannot replace molecular testing, it is a useful adjunct which requires minimal tissue and identifies false negative molecular results which occurred in 5% of our patients with eventually confirmed EGFR mutations.

Loss of ARID1A expression in colorectal carcinoma is strongly associated with mismatch repair deficiency.

ARID1A is a tumor suppressor gene involved in chromatin remodelling. ARID1A mutations and loss of protein expression occur commonly in endometrioid and gynecological clear cell carcinoma where they are associated with mismatch repair (MMR) deficiency. We assessed ARID1A expression in a large cohort of colorectal carcinomas (CRCs). Immunohistochemistry for ARID1A was performed on whole sections from 100 CRCs and on 1876 CRCs in tissue microarray format. There was complete concordance between the staining on whole slides and tissue microarray sections. Loss of staining was found in 110 (5.9%) of 1876 CRCs and was strongly associated with older age, right sided location, large size, BRAF V600E mutation, MMR deficiency, high histological grade and medullary morphology, (all P < .01). There was a trend towards loss of expression being more common in females (P = .06). When subclassified by combined BRAF V600E mutation and MMR status, loss of ARID1A expression was found most commonly in CRCs with the BRAF V600E mutated, MMR- deficient phenotype (58 of 232 cases, 25%, P < .01). In univariate and multivariate analysis, loss of ARID1A expression was not associated with overall survival-hazard ratio 1.05 (0.68-1.64) and 0.60 (0.24-1.44), respectively. All carcinomas arising in patients with known Lynch syndrome (n = 12) were ARID1A positive. We conclude that loss of ARID1A expression occurs in a small but significant proportion of CRCs where it is strongly correlated with mismatch repair deficiency and other clinical and pathological features associated with somatic hypermethylation.

Reflex ALK immunohistochemistry is feasible and highly specific for ALK gene rearrangements in lung cancer.

Fluorescence in situ hybridisation (FISH) is considered the gold standard for the detection of ALK gene rearrangements in lung adenocarcinoma. The presence of ALK gene rearrangement predicts response to specific targeted therapy, but these rearrangements are relatively rare and FISH studies are expensive, not widely available, potentially challenging to interpret and therefore difficult to undertake in all patients with non-small cell lung cancer. We developed and then deployed into the routine clinical setting a screening program for ALK gene rearrangement in all non-small cell lung cancer patients based on immunohistochemistry (IHC) with a mouse monoclonal antibody (clone 5A4).ALK IHC was strongly positive in 12 (4%) of 307 tumours from consecutive patients. Only 10 of these cancers were initially thought to be rearranged by diagnostic FISH studies. The two tumours which were IHC positive but initially interpreted as FISH negative underwent repeat FISH testing because of the discrepancy. Repeat FISH testing confirmed the presence of ALK gene rearrangement with the discrepancy being attributable to an atypical FISH pattern.Therefore, in our experienced hands, IHC for ALK performed on initial diagnosis of lung cancer is 100% specific for the presence of ALK gene rearrangement. When ALK IHC and FISH studies are discrepant, IHC may outperform FISH. Although our study was not intended to formally assess the sensitivity of ALK IHC, the 4% rate of gene rearrangements identified by this approach is consistent with the expected incidence in our population.We conclude that reflex ALK IHC followed by confirmatory FISH testing can be readily integrated into the routine clinical setting and represents a cost effective and practical approach to screening for these clinically significant gene rearrangements.

Immunohistochemistry for myc predicts survival in colorectal cancer.

MYC over-expression as determined by molecular means has been reported as a favorable prognostic biomarker in colorectal carcinoma (CRC). However MYC expression analysis is not available in the routine clinical setting. We investigated whether immunohistochemistry (IHC) for the myc protein using a novel commercially available rabbit monoclonal antibody [clone Y69] which is currently in widespread clinical use for lymphoma diagnosis could be used to predict outcome in resected CRC. Myc IHC was performed on a tissue microarray (TMA) comprising a retrospective cohort of 1421 CRC patients and scored blinded as to all clinical and pathological data. IHC was also performed on a subcohort of whole section CRCs to assess staining characteristics and concordance with TMA expression. MYC over-expression was found in 980 (69%) of CRCs and was associated with tumor stage and DNA mismatch repair/BRAF status. There was substantial agreement between TMA and whole section myc IHC (kappa = 0.742, p<0.01). CRCs with MYC over-expression demonstrated improved 5-year survival (93.2% vs. 57.3%), with the effect significantly modulated by the dominant effect of tumor stage, age at diagnosis and lymphovascular space invasion status on survival. We conclude that myc status as determined by IHC alone can be used to predict overall survival in patients with CRC undergoing surgical resection.

Melanoma arising from a long-standing pigmented trichoblastoma: clinicopathologic study with complementary aCGH/mutational analysis.

Trichoblastoma is a benign cutaneous adnexal tumor, composed mostly of follicular germinative cells. Its pigmented variant is colonized by numerous dendritic melanocytes. So far, only one case in the literature describes a combination of trichoblastoma and melanoma. We report the case of a 62-year-old man who had a slow-growing mass of the left flank present since childhood. This 8-cm mass was surgically removed when it became ulcerated and associated with axillary lymph nodes. Histologically, this tumor was strictly dermal and composed of 2 intermingled components. Large sheets of atypical, proliferating epithelioid cells predominated. Dispersed solid nests or cribriform epithelial islets encased in fibrous tissue were also seen. Some nests displayed a massive colonization by pigmented dendritic melanocytes. On immunohistochemical staining, the sheets of atypical cells expressed focally but strongly S100 protein, MelanA, HMB45, and MiTF. Epithelial structures diffusely expressed pancytokeratin AE1/AE3, KL1, and pleckstrin homology-like domain, family A, member 1. Based on these results, we diagnosed an intradermal melanoma, possibly developed from dendritic melanocytes colonizing a giant pigmented trichoblastoma. Direct sequencing of the melanoma revealed a rarely described NRAS mutation c.34G>T (G12C). Array comparative genomic hybridization displayed a complex profile somewhat divergent from standard melanoma profiles. The patient died of widespread metastatic disease 8 months after initial diagnosis.

Phosphaturic mesenchymal tumors show positive staining for somatostatin receptor 2A (SSTR2A).

Tumor-induced osteomalacia (TIO) is a paraneoplastic syndrome associated with tumors that secrete phosphaturic hormones, most notably fibroblast growth factor 23 (FGF23). The majority of tumors associated with this syndrome show stereotypical histological features and are now known as phosphaturic mesenchymal tumors (PMTs). We postulated that immunohistochemistry for somatostatin receptor 2A (SSTR2A) could be used to definitively identify PMTs or other tumors that cause TIO. Immunohistochemistry for FGF23 and SSTR2A was performed on 15 tumors from 14 patients with a definite diagnosis of TIO. All showed positive staining for both markers. While FGF23 staining was quite focal in some tumors, SSTR2A showed diffuse strong expression. In 40 control tumors not known to be associated with the clinical or biochemical features of TIO, FGF23 expression was found in 2 cases (one aneurysmal bone cyst and one osteosarcoma). SSTR2A expression was found in 9 control tumors (4 synovial sarcomas, 2 hemangiomas, 2 aneurysmal bone cysts and one osteosarcoma). Only one tumor (an aneurysmal bone cyst) showed positive staining for both FGF23 and SSTR2A. SSTR2A also commonly stained neoplastic and non-neoplastic endothelial cells. We conclude that neither FGF23 nor SSTR2A expression are specific for the diagnosis of PMT. However both stains are highly sensitive. Because of its diffuse strong expression and widespread availability, immunohistochemistry for SSTR2A is useful to confirm the diagnosis of PMT in an appropriate setting particularly if material is limited. Negative staining can serve as an excellent rule out test for this diagnosis.

Super selective radio embolization of the porcine kidney with 90yttrium resin microspheres: a feasibility, safety and dose ranging study.

PURPOSE: We determined whether super selective radio embolization of the porcine kidney was technically feasible and evaluated histopathological changes in the treatment target zone (upper or lower renal pole), adjacent nontargeted kidney, and adjacent and distant organs after administering (90)Y labeled vs bland resin microspheres. MATERIALS AND METHODS: We performed super selective radio embolization with (90)Y resin microspheres in 1 kidney and with an equivalent number of bland microspheres in the corresponding pole of the contralateral kidney as a control. The aim was to achieve radio embolization of a target zone equivalent to approximately a third of the kidney volume. A pathologist independently graded macroscopic and microscopic changes in the kidney, and adjacent and distant tissue resulting from incremental increases (0.15 to 0.35 GBq) in implanted activity in 6 pigs. RESULTS: We recorded grade 4 histological changes in the treatment target zone (upper or lower renal pole) in 5 of 6 pigs after injecting (90)Y resin microspheres with evidence of nephron sparing effects in the adjacent renal tissue at the lowest activity. At activity greater than 0.3 GBq increasing damage was noted to adjacent renal tissue beyond the treatment target zone. No toxicity was evident in adjacent or distant organs. CONCLUSIONS: Delivery of highly targeted intra-arterial radiotherapy to the kidney is feasible and safe in the pig model. Further evaluation is warranted as a potential treatment for advanced renal cell carcinoma or for localized disease in patients who are not candidates for surgery.