PDF file - 225K, Supplementary Figure 1. (A) Gender differences in frequency of EGFR mutations by stage. Men with EGFR mutations presented at the late stage more often than women (118/170, 69% vs. 235/423, 56%; p=0.002); women predominated at stage I (31% vs. 19%, p=0.004). (B) Stage distribution of EGFR exon 19 del and EGFR L858R. EGFR L858R were significantly more frequent at stage I than exon 19 del (83/246, 34% vs. 82/347, 24%; p=0.009); Fisher exact test, P value <0.01 is considered significant. Supplementary Figure 2: (A) Age and (B) stage distribution of KRAS mutations (4 major subtypes) Supplementary Figure 3: EGFR mutation nomogram ROC curve in the validation cohort. Supplementary Table 1. (A) The association of the smoking-free years and pack-years of smoking with EGFR mutational status. Smoking-free years impact on likelihood of EGFR mutation. (Multivariate logistic regression analysis). (B) The association of the smoking-free years and pack-years of smoking with KRAS mutational status. Smoking-free years do not impact on likelihood of KRAS mutation. (Multivariate logistic regression analysis). Supplementary Table 2. Age and gender differences in pack-years of smoking in relation to KRAS mutation nucleotide change. Women with G>T transversions had smoked less (average 34 pack-years vs. 40 pack-years, p=0.001) and were younger than men with the same nucleotide change (median age 64 vs. 67, p=0.006). Fisher exact test, P value <0.01 is considered significant.

PDF file - 308K, Complete list of target sequences for NanoString probes

PDF file - 76K, Examples of negative results with the NanoString assay for kinase fusions.

PDF file - 32K, Clinicopathological characteristics of study cohort

Myelodysplastic syndromes (MDS) are heterogeneous groups of clonal myeloid disorders characterized by unexplained persistent peripheral blood (PB) cytopenia(s) of one or more of the hematopoietic lineages, or bone marrow (BM) morphologic dysplasia in hematopoietic cells, recurrent genetic abnormalities, and an increased risk of progression to acute myeloid leukemia (AML). In the past several years, diagnostic, prognostic, and therapeutic approaches have substantially improved with the development of Next Generation Sequencing (NGS) diagnostic testing and new medications. However, there is no single diagnostic parameter specific for MDS, and correlations with clinical information, and laboratory test findings are needed to reach the diagnosis.

Primary neuroendocrine carcinoma of the breast is a rare variant, accounting for only 2% to 5% of diagnosed breast cancers, and may have relatively aggressive behavior. Mutational profiling of invasive ductal breast cancers has yielded potential targets for directed cancer therapy, yet most studies have not included neuroendocrine carcinomas. In a tissue microarray screen, we found a 2.4% prevalence (9/372) of neuroendocrine breast carcinoma, including several with lobular morphology. We then screened primary or metastatic neuroendocrine breast carcinomas (excluding papillary and mucinous) for mutations in common cancer genes using polymerase chain reaction-mass spectroscopy (643 hotspot mutations across 53 genes), or semiconductor-based next-generation sequencing analysis (37 genes). Mutations were identified in 5 of 15 tumors, including 3 with PIK3CA exon 9 E542K mutations, 2 of which also harbored point mutations in FGFR family members (FGFR1 P126S, FGFR4 V550M). Single mutations were found in each of KDR (A1065T) and HRAS (G12A). PIK3CA mutations are common in other types of breast carcinoma. However, FGFR and RAS family mutations are exceedingly rare in the breast cancer literature. Likewise, activating mutations in the receptor tyrosine kinase KDR (VEGFR2) have been reported in angiosarcomas and non-small cell lung cancers; the KDR A1065T mutation is reported to be sensitive to VEGFR kinase inhibitors, and fibroblast growth factor receptor inhibitors are in trials. Our findings demonstrate the utility of broad-based genotyping in the study of rare tumors such as neuroendocrine breast cancer.

Histologic classification of ampullary carcinomas into intestinal, pancreatobiliary, or other subtypes is easily achievable in some cases but difficult in others. Immunohistochemical (IHC) stains may allow distinction between the subtypes; however, their added value to routine hematoxylin and eosin (H&E) evaluation has not been systematically evaluated. Inconsistent histologic subtyping has hampered current clinical research and therapeutic trials. In this study, a consecutive series of 105 ampullary carcinomas was subtyped first by H&E evaluation and then by the evaluation of an IHC panel composed of CK7, CK20, CDX2, MUC1, and MUC2, and the added value of IHC was analyzed. By H&E, a consensus diagnosis, defined as concordant subtyping among at least 3 of the 4 independent study pathologists, was achieved in 81 of the 105 (77%) cases. There was excellent agreement for poorly differentiated and mucinous subtypes (κ=0.72 and 0.89, respectively) but only good agreement for intestinal and pancreatobiliary subtypes (κ=0.57 and 0.48, respectively) and poor agreement for mixed subtype (κ=0.09). By IHC, CK7 showed no informative value (being positive in ≥70% of the cases in both intestinal and pancreatobiliary subtypes), whereas a subtyping schema incorporating the combination staining patterns of CK20, CDX2, MUC1, and MUC2 did. By this schema, “intestinal subtype” was defined as having (1) positive staining for CK20 or CDX2 or MUC2 and negative staining for MUC1, or (2) positive staining for CK20, CDX2, and MUC2, irrespective of the MUC1 result; and “pancreatobiliary subtype” was defined as having positive staining for MUC1 and negative staining for CDX2 and MUC2, irrespective of CK20 results. Cases not fitting one of these 3 categories were regarded as “ambiguous” immunohistochemically. By combining this schema with H&E evaluation, 97 of the 105 cases (92%) could be classified into either intestinal or pancreatobiliary subtype. In particular, immunophenotyping allowed categorization of 75% of poorly differentiated adenocarcinomas and 69% of cases with mixed histologic features as either intestinal or pancreatobiliary subtype. Most mucinous adenocarcinomas (88%) were clearly intestinal subtype by IHC. Thus, our IHC schema enhanced the subtyping of ampullary carcinoma and, in combination with H&E evaluation, allowed a dichotomous classification in 92% of the cases. Should further independent studies reaffirm our findings, this schema may serve as a valuable tool in both diagnostic and research settings.

KRAS mutations define a clinically distinct subgroup of lung adenocarcinoma patients, characterized by smoking history, resistance to EGFR-targeted therapies, and adverse prognosis. Whether KRAS-mutated lung adenocarcinomas also have distinct histopathological features is not well established. We tested 180 resected lung adenocarcinomas for KRAS and EGFR mutations by high-sensitivity mass spectrometry-based genotyping (Sequenom) and PCR-based sizing assays. All tumors were assessed for the proportion of standard histological patterns (lepidic, acinar, papillary, micropapillary, solid, and mucinous), several other histological and clinical parameters, and TTF-1 expression by immunohistochemistry. Among 180 carcinomas, 63 (35%) had KRAS mutations (KRAS+), 35 (19%) had EGFR mutations (EGFR+), and 82 (46%) had neither mutation (KRAS−/EGFR−). Solid growth pattern was significantly over-represented in KRAS+ carcinomas: the mean±s.d. for the amount of solid pattern in KRAS+ carcinomas was 27±34% compared with 3±10% in EGFR+ (P<0.001) and 15±27% in KRAS−/EGFR− (P=0.033) tumors. Furthermore, at least focal (≥20%) solid component was more common in KRAS+ (28/63; 44%) compared with EGFR+ (2/35; 6%; P<0.001) and KRAS−/EGFR− (21/82; 26%; P=0.022) carcinomas. KRAS mutations were also over-represented in mucinous carcinomas and were significantly associated with the presence of tumor-infiltrating leukocytes and heavier smoking history. EGFR mutations were associated with non-mucinous non-solid patterns, particularly lepidic and papillary, lack of necrosis, lack of cytological atypia, hobnail cytology, TTF-1 expression, and never/light smoking history. In conclusion, extended molecular and clinicopathological analysis of lung adenocarcinomas reveals a novel association of KRAS mutations with solid histology and tumor-infiltrating inflammatory cells and expands on several previously recognized morphological and clinical associations of KRAS and EGFR mutations. Solid growth pattern was recently shown to be a strong predictor of aggressive behavior in lung adenocarcinomas, which may underlie the unfavorable prognosis associated with KRAS mutations in these tumors.

The phosphatidylinositol-3-kinase pathway is one of the most commonly altered molecular pathways in invasive breast carcinoma, with phosphatidylinositol-3-kinase catalytic subunit (PIK3CA) mutations in 25% of invasive carcinomas. Ductal carcinoma in situ (DCIS), benign papillomas, and small numbers of columnar cell lesions harbor an analogous spectrum of PIK3CA and AKT1 mutations, yet there is little data on usual ductal hyperplasia and atypical ductal and lobular neoplasias. We screened 192 formalin-fixed paraffin-embedded breast lesions from 75 patients for point mutations using a multiplexed panel encompassing 643 point mutations across 53 genes, including 58 PIK3CA substitutions. PIK3CA point mutations were identified in 31/62 (50%) proliferative lesions (usual ductal hyperplasia and columnar cell change), 10/14 (71%) atypical hyperplasias (atypical ductal hyperplasia and flat epithelial atypia), 7/16 (44%) lobular neoplasias (atypical lobular hyperplasia and lobular carcinoma in situ), 10/21 (48%) DCIS, and 13/37 (35%) invasive carcinomas. In genotyping multiple lesions of different stage from the same patient/specimen, we found considerable heterogeneity; most notably, in 12 specimens the proliferative lesion was PIK3CA mutant but the concurrent carcinoma was wild type. In 11 additional specimens, proliferative epithelium and cancer contained different point mutations. The frequently discordant genotypes of usual ductal hyperplasia/columnar cell change and concurrent carcinoma support a role for PIK3CA-activating point mutations in breast epithelial proliferation, perhaps more so than transformation. Further, these data suggest that proliferative breast lesions are heterogeneous and may represent non-obligate precursors of invasive carcinoma.

Micropapillary carcinoma of the breast is associated with increased rates of lymph node metastasis and lymphovascular invasion. While activating point mutations in PIK3CA (encoding phosphatidylinositol-3-kinase catalytic subunit) or AKT1 are found in 25% to 30% of invasive ductal carcinomas, the mutational profile of invasive micropapillary carcinomas has not been characterized in detail. Micropapillary carcinomas, concurrent metastatic and precursor breast lesions from 19 patients were identified. Lesional tissue was punched from paraffin-tissue blocks, and genomic DNA was extracted and screened for a large panel of known hotspot mutations using multiplex polymerase chain reaction and mass-spectroscopy analysis (643 mutations in 53 genes). Hotspot point mutations were identified in 35% (7/20) of micropapillary breast carcinomas, including PIK3CA exons 7, 9 and 20 hotspots, as well as the AKT1 plekstrin homology domain mutation (E17K); mutations in TP53 and KRAS were each found in a single patient. In 6 patients, micropapillary and non-micropapillary components of the same tumor were separately tested, yielding concordant results in five; one had a wild type micropapillary component, but a PIK3CA mutation in the invasive ductal component. Concurrent lymph node metastases were mostly wild type (2/8 mutant). Accompanying ductal carcinoma in situ had point mutations in 45% (5/11), mostly concordant with invasive carcinoma; however, mutational status of other breast proliferative lesions was generally discordant with accompanying carcinoma. The rate of PIK3CA mutations in this series of micropapillary carcinomas is similar to invasive ductal carcinomas; however, there may be an enrichment of AKT1 mutations (10%). The non-micropapillary components and precursor lesions occasionally had different mutations.

Primary neuroendocrine carcinoma of the breast is a rare variant, accounting for only 2% to 5% of diagnosed breast cancers, and may have relatively aggressive behavior. Mutational profiling of invasive ductal breast cancers has yielded potential targets for directed cancer therapy, yet most studies have not included neuroendocrine carcinomas. In a tissue microarray screen, we found a 2.4% prevalence (9/372) of neuroendocrine breast carcinoma, including several with lobular morphology. We then screened primary or metastatic neuroendocrine breast carcinomas (excluding papillary and mucinous) for mutations in common cancer genes using polymerase chain reaction-mass spectroscopy (643 hotspot mutations across 53 genes), or semiconductor-based next-generation sequencing analysis (37 genes). Mutations were identified in 5 of 15 tumors, including 3 with PIK3CA exon 9 E542K mutations, 2 of which also harbored point mutations in FGFR family members (FGFR1 P126S, FGFR4 V550M). Single mutations were found in each of KDR (A1065T) and HRAS (G12A). PIK3CA mutations are common in other types of breast carcinoma. However, FGFR and RAS family mutations are exceedingly rare in the breast cancer literature. Likewise, activating mutations in the receptor tyrosine kinase KDR (VEGFR2) have been reported in angiosarcomas and non-small cell lung cancers; the KDR A1065T mutation is reported to be sensitive to VEGFR kinase inhibitors, and fibroblast growth factor receptor inhibitors are in trials. Our findings demonstrate the utility of broad-based genotyping in the study of rare tumors such as neuroendocrine breast cancer.

The pathognomonic genetic alteration in chronic myeloid Leukemia is the formation of the BCR-ABL1 fusion gene, which produces a constitutively active tyrosine kinase that drives leukemic transformation. Targeted tyrosine kinase inhibitor treatment with imatinib, nilotinib, dasatinib, bosutinib, and ponatinib is the cornerstone of modern therapy for this hematologic malignancy. Real-time quantitative RT-PCR (RT-qPCR, also RQ-PCR) of BCR-ABL1 RNA is a necessary laboratory technique for monitoring the efficacy of tyrosine kinase inhibitor therapy and quantitatively assessing minimal residual disease. The molecular response measured by BCR-ABL1 RT-qPCR assists in identifying suboptimal responses and can help inform the decision to switch to alternative therapies that may be more efficacious (or to pursue more stringent monitoring). Furthermore, the tyrosine kinase inhibitor-mediated molecular response provides valuable risk stratification and prognostic information on long-term outcomes. Despite these attributes, informed, universal, practical utilization of this well-established monitoring test will require heightened efforts by the molecular diagnostics laboratory community to adopt the standardized reporting units of the International Scale. Without widespread adoption of the International Scale, the consensus major molecular response and early molecular response treatment thresholds will not be definable, and optimal clinical outcomes for patients with chronic myeloid leukemia may not be achieved.

Somatic mutations in PIK3CA are commonly seen in invasive breast cancer and several other carcinomas, occurring in three hotspots: codons 542 and 545 of exon 9 and in codon 1047 of exon 20. We designed a locked nucleic acid (LNA)-PCR sequencing assay to detect Low Levels of mutant PIK3CA DNA with attention to avoiding amplification of a pseudogene on chromosome 22 that has >95% homology to exon 9 of PIK3CA. We tested 60 FFPE breast DNA samples with known PIK3CA mutation status (48 cases had one or more PIK3CA mutations, and 12 were wild type) as identified by PCR-mass spectrometry. PIK3CA exons 9 and 20 were amplified in the presence or absence of LNA-oligonucleotides designed to bind to the wild-type sequences for codons 542, 545, and 1047, and partially suppress their amplification. LNA-PCR sequencing confirmed all 51 PIK3CA mutations; however, the mutation detection rate by standard Sanger sequencing was only 69% (35 of 51). Of the 12 PIK3CA wild-type cases, LNA-PCR sequencing detected three additional H1047R mutations in "normal" breast tissue and one E545K in usual ductal. hyperplasia. Histopathological review of these three normal breast specimens showed columnar cell change in two (both with known H1047R mutations) and apocrine metaplasia in one. The novel LNA-PCR shows higher sensitivity than standard Sanger sequencing and did not amplify the known pseudogene.

Activating point mutations in the phosphatidylinositol-3-kinase catalytic subunit (PIK3CA) are among the most common molecular defects in invasive breast cancer. Point mutations in the downstream kinase AKT1 are seen in a minority of carcinomas. These mutations are found preferentially in estrogen receptor–positive and Her2-positive breast carcinomas; however, special morphologic types of breast cancer have not been well studied. Twenty-nine cases of pure invasive mucinous carcinoma and 9 cases of ductal carcinoma with mucinous differentiation were screened for a panel of point mutations (>321 mutations in 30 genes) using a multiplex polymerase chain reaction panel with mass spectroscopy readout. In addition, associated ductal carcinoma in situ, hyperplasia, or columnar cell lesions were separately tested where available (25 lesions). In 3 invasive cases and 15 ductal carcinoma in situ/proliferative lesions, PIK3CA hotspot mutations were, instead, tested by direct sequencing. No point mutations were identified in invasive mucinous breast carcinoma. This contrasts with the 35% frequency of PIK3CA mutations in a comparative group of invasive ductal carcinomas of no special type. Interestingly, PIK3CA hotspot point mutations were identified in associated ductal carcinoma in situ (3/14) and hyperplasia (atypical ductal hyperplasia [2/3], usual ductal hyperplasia [2/3], columnar cell change [1/5]), suggesting that PIK3CA mutations may play a role in breast epithelial proliferation. This series represents the largest study, to date, of PIK3CA genotyping in mucinous carcinoma and supports the unique pathogenetics of invasive mucinous breast carcinoma.

Background EGFR mutation status is the best predictor of response to tyrosine kinase inhibitors (TKIS) in primary lung adenocarcinoma. Approximately 70% of lung cancers are diagnosed in advanced stages where small biopsies and cytological specimens are the only source of material for both diagnosis and mutation testing. Specific antibodies that can detect mutant EGFR protein were evaluated for the detection of EGFR mutation by immunohistochemistry (IHC) in cytology and small biopsy specimens. Methods Assessment of EGFR mutation status was performed by using antibodies specific to the two major forms of mutant EGFR, exon 21 L858R and exon 19 deletion (15 bp). The study was performed in 145 lung adenocarcinomas, including cytology material, core biopsy, and decalcified bone biopsy. Stains were scored as negative (0), 1+ (weak and focal), 2+ (moderate intensity and focal), and 3+ (strong and diffuse). The result of the IHC stains was correlated with mutations status determined by standard molecular methods. Results Validation using clinical material showed deletions in exon 19 were detected in 35% and L858R mutation in 17.6% of all cases by standard molecular methods. A cutoff value of 2+ was used as positive by IHC. No wild type cases were immunoreactive. The positive predictive value (PPV) and specificity for both antibodies was 100%. The antibodies performed well in cytology, core biopsies and decalcified bone biopsies. Conclusion Immunostaining to detect specific mutant EGFR shows a good correlation with mutation analysis and can be used as a screening method to identify patients for TKI therapy. IHC methodology is potentially useful when molecular analysis is not available and for use in small biopsies when material is too scant for molecular tests. Importantly mutation specific antibodies are useful in determining EGFR status in tissues obtained from bone biopsy as decalcification processes used in molecular based studies often result in DNA degradation hindering mutation detection.

Morphologic "special types" of breast carcinomas have been recognized for many years, and their molecular and genetic properties have not been specifically studied until recently. Lobular carcinoma lacks functional E-cadherin expression but shares molecular similarities with low-grade invasive ductal carcinomas. Papillary carcinoma is relatively rare, and molecular features are just being elucidated. We report a case of concurrent invasive lobular and papillary carcinoma, the latter with extensive nodal involvement. Multiplex screening for activating point mutations identified different point mutations in the distinct morphologic components: lobular PIK3CA H1047R, papillary; PIK3CA Q546P, and IDH1 R132H. These molecular data favor coincidental "collision tumors" over clonal evolution. The IDH1 R132H point mutation is common in gliomas and acute myelogenous leukemia, but this has not been previously reported in breast carcinoma. The characterization of activating point mutations in morphologic special types of breast carcinoma may suggest avenues amenable to targeted therapy.

Abstract Purpose: The molecular epidemiology of most EGFR and KRAS mutations in lung cancer remains unclear. Experimental Design: We genotyped 3,026 lung adenocarcinomas for the major EGFR (exon 19 deletions and L858R) and KRAS (G12, G13) mutations and examined correlations with demographic, clinical, and smoking history data. Results:EGFR mutations were found in 43% of never smokers and in 11% of smokers. KRAS mutations occurred in 34% of smokers and in 6% of never smokers. In patients with smoking histories up to 10 pack-years, EGFR predominated over KRAS. Among former smokers with lung cancer, multivariate analysis showed that, independent of pack-years, increasing smoking-free years raise the likelihood of EGFR mutation. Never smokers were more likely than smokers to have KRAS G > A transition mutation (mostly G12D; 58% vs. 20%, P = 0.0001). KRAS G12C, the most common G > T transversion mutation in smokers, was more frequent in women (P = 0.007) and these women were younger than men with the same mutation (median 65 vs. 69, P = 0.0008) and had smoked less. Conclusions: The distinct types of KRAS mutations in smokers versus never smokers suggest that most KRAS-mutant lung cancers in never smokers are not due to second-hand smoke exposure. The higher frequency of KRAS G12C in women, their younger age, and lesser smoking history together support a heightened susceptibility to tobacco carcinogens. Clin Cancer Res; 18(22); 6169–77. ©2012 AACR.