Detection of CALR Mutation in Clonal and Nonclonal Hematologic Diseases Using Fragment Analysis and Next-Generation Sequencing

Detection of CALR Mutation in Clonal and Nonclonal Hematologic Diseases Using Fragment Analysis and Next-Generation Sequencing Juli-Anne Gardner, MD, 1 Jason D. Peterson, MS, 1 Scott A. Turner, PhD, 1 Barbara L. Soares, 2 Courtney R. Lancor, 1 Luciana L. dos Santos, PhD, 2 Prabhjot Kaur, MD, 1 Deborah L. Ornstein, MD, 1 Gregory J. Tsongalis, PhD, 1 and Francine B. de Abreu, PhD 1 From the 1 Department of Pathology and Laboratory Medicine, Geisel School of Medicine at Dartmouth, Hanover, NH, and Dartmouth Hitchcock Medical Center and Norris Cotton Cancer Center, Lebanon, NH, and 2 Universidade Federal De Sao Joao Del Rei, Divinopolis, Minas Gerais, Brazil Key Words: Clonal diseases; MPN; CALR gene; Somatic mutation; Fragment analysis; Sanger sequencing; Next-generation sequencing; Routine clinical laboratory Am J Clin Pathol October 2016;146:448-455 DOI: 10.1093/AJCP/AQW129 ABSTRACT Objectives: To describe three methods used to screen for frameshift mutations in exon 9 of the CALR gene. Methods: Genomic DNA from 47 patients was extracted from peripheral blood and bone marrow using the EZ1 DNA Blood Kit (Qiagen, Valencia, CA) and quantified by the Quant-iT PicoGreen dsdna Assay Kit (Invitrogen, San Diego, CA). After clinical history, cytogenetics, and molecular tests, patients were diagnosed with either clonal or nonclonal hematologic diseases. CALR screening was primarily performed using fragment analysis polymerase chain reaction, then next-generation sequencing and Sanger sequencing. Results: Among the 18 patients diagnosed with clonal diseases, one had acute myeloid leukemia (positive for trisomy 8), and 17 had myeloproliferative neoplasms (MPNs), including chronic myeloid leukemia (CML), essential thrombocythemia (ET), primary myelofibrosis (PMF), and polycythemia vera (PV). Patients with CML were positive for the BCR-ABL1 fusion. Ten patients were positive for JAK2 (PMF, n ¼ 1; ET, n ¼ 2; PV, n ¼ 7), and three were CALR positive (ET, n ¼ 1; PMF, n ¼ 2). Patients diagnosed with a nonclonal disease were negative for JAK2, BCR-ABL, and CALR mutations. Conclusions: Screening for CALR mutations is essential in BCR-ABL negative MPNs since it not only provides valuable diagnostic and prognostic information but also identifies potential treatment targets. Since this study describes the importance of screening for known and novel biomarkers, we described in detail three methods that could be easily integrated into a clinical laboratory. Clonal hematologic proliferations are derived from a single progenitor cell with the same chromosomal or molecular abnormality. Nonclonal or reactive hematologic proliferations, by definition, lack a known associated chromosomal or molecular aberrancy. As described in the current World Health Organization (WHO) classification, diagnosis of clonal and nonclonal/reactive proliferations requires the integration of clinical history, which includes cytologic and histomorphology, immunohistochemistry, cytogenetics, and molecular genetic studies. Clinical findings often overlap in these conditions, and the distinction can be especially difficult in those cases with cytologic and/ or histomorphologic findings suggestive of a clonal proliferation but lacking an associated chromosomal or molecular abnormality. Successful distinction between the two conditions is paramount as clonal proliferations require continued clinical monitoring at a minimum and may require aggressive therapy. Nonclonal or reactive proliferations may require continued evaluation to determine the underlying cause of the reactive process or may simply be monitored clinically. Myeloid malignancies are clonal diseases that comprise chronic neoplasms such as myeloproliferative neoplasms (MPNs), myelodysplastic syndromes, chronic myelomonocytic leukemia, and acute myeloid leukemia (AML). The MPNs are chronic myeloid disorders of hematopoietic stem cells characterized by proliferation of one or more myeloid lineages that retain the ability to terminally differentiate into mature cells, causing cytoses in the peripheral blood. Chronic myeloid leukemia (CML) with the BCR-ABL1 fusion gene is the prototype disease for this category of 448 Am J Clin Pathol 2016;146:448-455 American Society for Clinical Pathology, 2016. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

neoplasms. The three most common BCR-ABL1 negative MPNs are essential thrombocythemia (ET), polycythemia vera (PV), and primary myelofibrosis (PMF). JAK2 mutation analysis is critical in the diagnostic algorithm of BCR- ABL negative MPNs. 1-4 Mutation status can help guide treatment and has prognostic implications. JAK2 V617F (c.1849g > T; p.v617f) mutations are found in most patients with PV. 5,6 In patients with ET or PMF, JAK2 V617F mutation is seen in 50% to 60%. 7 However, 10% of patients with ET and PMF who are JAK2 negative show an activation mutation in the MPL gene. 8,9 The natural history of BCR-ABL1 negative MPNs is characterized by thromboembolic events, progression to myelofibrosis, and AML. 3 Thus, the importance of identifying gene mutations for not only diagnosis but also prognostic reasons is critical. Equally important is identifying those patients with presentations that mimic clonal neoplasms but are reactive in nature since these patients may require continued evaluation for the underlying cause of their reactive cytosis or may simply be followed clinically and not require treatment. JAK2-negative PMF and ET have long presented a diagnostic dilemma for pathologists and clinicians. Recently, INDELs (insertions and deletions) resulting in frameshift mutations in exon 9 of the calreticulin (CALR) gene were described in patients with PMF and ET negative for JAK2 and MPL mutations. 2,10 Data from these two studies provide strong evidence that CALR mutations act as a novel driver mutation early in the pathogenesis of these disorders with a functional role similar to JAK2 and MPL. Thus far, CALR mutations seen in ET and PMF involve either somatic insertions or deletions in exon 9. CALR mutations are associated with a distinct clinical phenotype as well as different patterns of complications and differences in overall survival. 1,2,11-13 Discovery of CALR mutations has enabled the proper diagnosis of a subset of JAK2- and MPL-negative ET and PMF, providing definitive evidence of clonality and identifying potential treatment options. The aim of this study was to describe in detail three different methods used to detect frameshift mutations in exon 9 of the CALR gene in routine clinical laboratories, including fragment analysis polymerase chain reaction (PCR), nextgeneration sequencing (NGS), and Sanger sequencing. Figure 1. Genomic DNA from all patients was extracted from peripheral blood and bone marrow using the EZ1 DNA Blood Kit (Qiagen, Valencia, CA) and quantified by the Quant-iT PicoGreen dsdna Assay Kit (Invitrogen, San Diego, CA). The clonal diseases were composed of CML (n ¼ 4), AML (n ¼ 1), and MPNs (ET, n ¼ 3; PMF, n ¼ 3; PV, n ¼ 7) and classified based on the current WHO classification, and the nonclonal diseases were categorized as erythrocytosis (n ¼ 15), leukocytosis (n ¼ 8), and thrombocytosis (n ¼ 6) if they failed to meet criteria for an MPN as outlined in the current WHO classification (Figure 1). All 47 samples were screened for JAK2 and CALR mutations, and 35 were also screened for the BCR-ABL1 fusion. CALR Fragment Analysis All 47 samples were screened for CALR mutations using a clinical laboratory-validated CALR fragment analysis assay. A set of primers flanking exon 9 of the CALR gene was used to detect frameshift mutations. 10 The conventional PCR was performed using 0.5 mm each primer, 1 AmpliTaq Gold 360 Master Mix (Applied Biosystems, Foster City, CA), 10 ng DNA, and nuclease-free water (total reaction volume 25 ml). PCR conditions were as described by Klampfl et al. 10 Human Genomic DNA Male (Promega, Madison, WI) was used as a negative control in each reaction. PCR products were analyzed using the ABI 3500 Genetic Analyzer with POP-7 polymer and 50-cm capillary. The final results were analyzed using GeneMapper 4.1 Software (Applied Biosystems). A sample was considered negative when a wild-type (wt) allele peak was observed at 263 base pairs (bp) and positive when an extra allele peak was observed in conjunction to the wt peak Figure 2. Materials and Methods Samples This study includes 47 patients diagnosed with either clonal (n ¼ 18) or nonclonal (n ¼ 29) hematologic disease according to the current WHO classification, which involves clinical history, cytologic and histomorphology, immunohistochemistry, cytogenetics, and molecular tests Figure 1 Total of clonal and nonclonal samples and their mutation status. American Society for Clinical Pathology Am J Clin Pathol 2016;146:448-455 449 449

Gardner et al /DETECTION OF CALR MUTATION IN PATIENTS WITH HEMATOLOGIC DISEASES To integrate this test into routine clinical use, a validation was performed to evaluate the limit of detection (LOD), precision, and accuracy of the assay. LOD was assessed in terms of DNA input and allelic frequency. For precision, six samples (two positives and four negative) were screened in duplicate in four independent runs by two different operators. Accuracy was assessed by comparison of results to Sanger sequencing. A B CALR NGS All samples positive for frameshift mutations (type 1 mutation, 52-bp deletion; type 2 mutation, 5-bp insertion) and eight negative samples were sequenced using the Illumina TruSight Myeloid Sequencing Panel (Illumina, San Diego, CA) that consists of 54 genes and 568 amplicons. For library preparation, at least 50 ng genomic DNA was used for oligo hybridization to targeted regions, followed by extension and ligation. Indices and sequence adapters were added by PCR amplification. Finally, libraries were purified, normalized, quantified, pooled, and sequenced on the Illumina MiSeq System. The average cluster density was between 1,200 and 1,400 K/mm 2, and more than 90.0% of the clusters passed Q30. Following FASTQ file generation, sequences were aligned twice using the MiSeq Reporter Software (Illumina) and two independent alignment settings. These settings included both the default TruSight Myeloid Sequencing (Illumina) panel workflow parameters and alignment settings optimized for large INDELs. The optimized settings allow for the detection of large INDELs (the default maximum detectable INDEL length is 25 bp) (Figure 2). This maximum default size can be manually configured by modifying the CustomAmpliconAlignerMaxIndelSize parameter in the sample sheet when setting up a sequencing run. A value of 80 was used for this parameter to identify the CALR mutations. While setting this parameter to a higher value does increase sensitivity to larger INDELs, it also requires additional alignment time. Both the default and CALR-optimized analysis pipelines used the Burrows- Wheeler aligner to map reads to the reference genome (GRCh37/hg19). Illumina s proprietary somatic variant caller was used to identify variants in both the default and optimized alignment files. Variant call format (VCF) files were then uploaded to VariantStudio v2.2 (Illumina), where variants were annotated, classified, and filtered for quality (pass), read depth (>500), allelic frequency (>5%), significance, and population frequency (<1% minor allele frequency). For the large INDEL optimized VCF file, a custom filter C Figure 2 Sample negative for CALR frameshift mutation. A, CALR fragment analysis. The wild-type allele peak is observed at 263 base pairs. B, CALR next-generation sequencing. Default and optimized Integrative Genomics Viewer image of exon 9 of the CALR gene. The gray bars within the coverage track represent the depth of coverage at each locus; the red bars within the alignment track represent the reads mapped to the targeted region; the two black center lines flank the bases centered in the display screen. C, CALR Sanger sequencing. was also used in VariantStudio to include both variants that pass all filters and those that fail the strand bias filter. The validation process for the Myeloid Sequencing Panel included samples negative and positive for frameshift mutations (INDELs) in the CALR, FLT3, and NPM1 genes and point mutations in the FLT3, IDH2, JAK2, MET, NRAS, and TP53 genes. Each run also included an internal control (Tru-Q 4 [5% Tier], DNA Reference Standard) (Horizon 450 Am J Clin Pathol 2016;146:448-455 American Society for Clinical Pathology 450

Diagnostics, Cambridge, UK). Precision was assessed using a sample positive for type 1 mutation in the CALR gene, two samples negative for mutations in the CALR gene, two clinical samples, and an internal control. All samples were sequenced in two independent runs by different operators. Accuracy was assessed by comparison of results to another NGS panel. Following this initial validation, it was determined that our analytical pipeline could not reliably detect large INDELs. For this reason, at the time of the initial implementation of the sequencing assay, screening of frameshift mutations in the CALR gene was performed by fragment analysis. However, with the development of the optimized analysis pipeline, samples positive for type 1 and type 2 mutations and six negative samples were sequenced and included in the validation process. CALR Sanger Sequencing To confirm the results obtained using the CALR fragment analysis and the CALR NGS, all samples positive for frameshift mutations (type 1 mutation, 52-bp deletion; type 2 mutation, 5-bp insertion) were screened using Sanger sequencing. In this assay (Figure 2), exon 9 was PCR amplified using 500 lmol/l for both forward and reverse primers and HotStar Plus Master Mix (Qiagen; 2 Qiagen PCR buffer, 3 mmol/l MgCl 2, 400 lmol/l each dntp). Cycling conditions were 95 C for 2 minutes, followed by 35 cycles at 95 C for 30 seconds, 60 C for 30 seconds, and 72 C for 30 seconds. All products were visualized on a 2% agarose gel confirming amplification of CALR mutations. Products were sequenced from the 5 0 and 3 0 directions using the BigDye Terminator v3.1 sequencing kit (LifeTechnologies, Carlsbad, CA) and purified using Performa DTR gel filtration cartridges (EdgeBio, Gaithersburg, MD) following the manufacturer s guidelines. Sanger sequencing was performed by capillary electrophoresis using the 3500 Genetic Analyzer (LifeTechnologies), and sequences were analyzed using the Variant Reporter v1.0 (Applied Biosystems) against a CALR refseq (NM_004343). Results JAK2, BCR-ABL, and CALR Mutation Screening Among the 18 clonal hematologic disease samples, 10 were positive for JAK2 mutation (V617F), four were BCR- ABL1 fusion positive, and three were positive for CALR mutation (type 1 mutation, 52-bp deletion; type 2 mutation, 5-bp insertion) (Figure 1). Another was negative for JAK2 and CALR mutations, and it was not screened for BCR-ABL1 fusion. Samples with clonal diseases were composed of AML (n ¼ 1) and MPNs (CML, n ¼ 4; ET, n ¼ 3; PMF, n ¼ 3; PV, n ¼ 7). A patient diagnosed with AML was negative for both JAK2 and CALR mutations (BCR-ABL1 was not performed). However, the patient was positive for trisomy 8. Four of 17 patients diagnosed with MPNs were positive for BCR-ABL1 fusion, 10 were positive for JAK2 (PMF, n ¼ 1; ET, n ¼ 2; PV, n ¼ 7), and three were CALR positive (ET, n ¼ 1; PMF, n ¼ 2) (Figure 1). All samples positive for CALR mutation were negative for JAK2 mutations. CALR mutation was not found in patients with PV Table 1. In the nonclonal hematologic disease group, all samples were negative for JAK2, BCR-ABL1, and CALR mutations (Table 1). CALR Fragment Analysis For the validation process of the CALR fragment analysis assay, analytical sensitivity was assessed in terms of DNA input and allelic frequency. DNA input was established by varying the amount of genomic DNA in the reaction (100 ng, 50 ng, 10 ng, 5 ng, 1 ng), and allelic frequency was established by serial dilution of a positive sample into a negative sample (100%, 50%, 25%, 12.5%, 5.0%, 2.5%). Since the intensity of the electropherograms decreased according to the amount of DNA and allelic frequency, the LOD established was 10 ng for DNA input and 5% for allelic frequency. Reproducibility or the lack of interassay/within-assay variability was demonstrated when identical calls (100% precision) were made for the six duplicate samples on four independent runs generated by different operators. One hundred percent accuracy was observed when identical calls were made for positive and negative samples characterized by Sanger sequencing. Among the 47 samples screened for frameshift mutations in the CALR gene using this assay, 44 were negative, two were positive for type 1 mutation (52-bp deletion), and one was positive for type 2 mutation (5-bp insertion). Figure 3A shows one sample negative for frameshift mutation (a unique wt allele at 263 bp), and Figure 4A shows samples positive for insertion and deletion, respectively. Both patients had an additional peak (mutant allele) in conjunction to the wt allele peak (263 bp). The patient in Figure 4A had a mutant allele peak at 268 bp, and a patient in had a mutant allele peak at 211 bp. CALR NGS Analysis Default and optimized alignment files from all samples were analyzed using VariantStudio v2.2. The sample with a type 2 mutation and the negative samples were correctly identified as positive and negative, respectively, using the VCF file generated by the default workflow parameters Figure 3B and Figure 4B. Samples with type 1 mutation (52-bp deletion) in the CALR gene were not identified as positive using the VCF file generated using the default American Society for Clinical Pathology Am J Clin Pathol 2016;146:448-455 451 451

Gardner et al /DETECTION OF CALR MUTATION IN PATIENTS WITH HEMATOLOGIC DISEASES Table 1 Mutation Status in Both Clonal and Nonclonal Groups Hematologic Disease workflow parameters. However, lower coverage within this region was observed using the Broad Institute s Integrative Genomics Viewer (IGV). IGV analysis also showed a small deletion and multiple point mutations across exon 9 in type 1 positive samples, but this profile was not visible in either the negative samples or the type 2 mutation sample (5-bp insertion). When using VCF files generated by the optimized workflow parameters, all 11 samples (negative, n ¼ 8; type 1 positive, n ¼ 2; type 2 positive, n ¼ 1) were identified as negatives and positives. IGV analysis of the negative sample and sample with the type 2 mutation showed the same results detected using the default workflow (Figure 3B and Figure 4B, respectively). However, for samples with type 1 mutations, IGV analysis confirmed the presence of the deletion (Figure 4B). CALR Sanger Sequencing Negative and positive samples previously screened for frameshift mutations in the CALR gene using fragment analysis and NGS were confirmed by Sanger sequencing, which showed 100% agreement Figure 3C and Figure 4C. Discussion Total No. of Samples The calreticulin protein can be found within the endoplasmic reticulum, where it acts as a chaperone protein and is involved in calcium homeostasis. 14,15 It can also be found in intracellular compartments and on the cell surface, where it has been implicated in several biological processes, including proliferation, apoptosis, and immune-mediated cell death. 16-19 Mutation Screening (No.) JAK2 BCR-ABL CALR Clonal 19 AML 1 Negative (1) N/P (1) Negative (1) CML 4 Negative (4) Positive (4) Negative (4) ET (MPN) 3 Negative (1) Negative (2) Negative (2) Positive (2) N/P (1) Positive (1) PMF (MPN) 3 Negative (2) Negative (1) Negative (1) Positive (1) Negative (2) Positive (2) PV (MPN) 7 Positive (7) Negative (6) Negative (7) N/P (1) Nonclonal 29 Erythrocytosis 15 Negative (15) Negative (8) Negative (15) N/P (7) Leukocytosis 8 Negative (8) Negative (6) Negative (8) N/P (2) Thrombocytosis 6 Negative (6) Negative (5) Negative (6) N/P (1) AML, acute myeloid leukemia; CML, chronic myeloid leukemia; ET, essential thrombocythemia; MPN, myeloproliferative neoplasm; N/P, screening not performed; PMF, primary myelofibrosis; PV, polycythemia vera. Mutations detected in the CALR gene are either insertions or deletions located in exon 9. Although these mutations are variable, they result in the development of one mutant protein with a novel C terminus that contains a number of positively charged amino acids. 20 Depending on the type of the mutation, the mutant protein retains varying amounts of the negatively charged amino acids of normal calreticulin, which may be associated with qualitatively different phenotypes. 10 Several mutation types have been reported in exon 9 of the CALR gene, but type 1 (52-bp deletion) and type 2 (5-bp insertion) are the most frequently identified. Type 1 mutation eliminates almost all negatively charged amino acids, and type 2 retain approximately half the negatively charged amino acids. 10 Mutations in the JAK2 and MPL genes were included in the WHO diagnostic criteria of 2008 for the three classic groups of MPNs (ET, PMF, and PV). However, the identification of diagnostic markers for BCR-ABL1 negative MPNs has recently advanced with the discovery of the CALR mutation. For this reason, it became critical to test for mutations in JAK2, MPL, and CALR genes in patients with BCR-ABL1 negative MPNs for management and prognosis. In this study, we demonstrated that CALR frameshift mutations can be detected using not only fragment analysis and Sanger sequencing but also NGS. There are numerous publications showing different mechanisms that detect frameshift mutations in the CALR gene, such as fragment analysis, high-resolution melting (HRM) analysis, TaqMan real-time PCR assay, and Sanger sequencing. 10,21-25 However, to our knowledge, this is the first study that describes a clinically validated assay using a targeted NGS panel. 452 Am J Clin Pathol 2016;146:448-455 American Society for Clinical Pathology 452

A A B B C Figure 3 Sample positive for type 2 mutation (5 base pair [bp] insertion) in the CALR gene. A, CALR fragment analysis. The wild-type allele peak is observed at 263 bp in conjunction to an extra allele peak at 268 bp. B, CALR next-generation sequencing (NGS). Default and optimized Integrative Genomics Viewer (IGV) image of exon 9 of the CALR gene. The gray bars within the coverage track represent the depth of coverage at each locus; the red bars within the alignment track represent the reads mapped to the targeted region; the two black center lines flank the bases centered in the display screen; the blue symbols represent the inserted nucleotides (identified on both IGV images generated by both NGS analysis). C, CALR Sanger sequencing. In this study, CALR mutations were identified only in patients with PMF and ET and were mutually exclusive with JAK2 mutations, which supports previous studies. 10,20,24,25 JAK2 mutations were seen in patients with PV, ET, and PMF, and those with PV had exclusively JAK2 C Figure 4 Sample negative for type 1 mutation (52 base pair [bp] deletion) in the CALR gene. A, CALR fragment analysis. The wild-type allele peak is observed at 263 bp in conjunction to an extra allele peak at 211 bp. B, CALR next-generation sequencing (NGS). Default and optimized Integrative Genomics Viewer (IGV) image of exon 9 of the CALR gene. The gray bars within the coverage track represent the depth of coverage at each locus; the red bars within the alignment track represent the reads mapped to the targeted region; the two black center lines flank the bases centered in the display screen; the black dash represents the deleted nucleotides (identified only on the second IGV image that represents the optimized NGS analysis). C, CALR Sanger sequencing. mutations. As expected, the BCR-ABL1 fusion gene was found in all patients with CML. Since patients with MPNs who are positive for CALR mutations are often younger and have a lower hemoglobin American Society for Clinical Pathology Am J Clin Pathol 2016;146:448-455 453 453

Gardner et al /DETECTION OF CALR MUTATION IN PATIENTS WITH HEMATOLOGIC DISEASES level and leukocyte count, higher platelet count, and longer overall survival than those with JAK2-mutated MPNs, the detection of CALR mutations has become a significant prognostic indicator, and CALR mutation status is also relevant for assessing clinical phenotype. For example, patients with ET who are CALR positive have a lower risk of thrombosis compared with patients with JAK2 mutations and also have a lower hemoglobin level and WBC count, as well as a higher platelet count. 11,26 Patients with PMF who harbor CALR mutations are younger, have a higher platelet count, have a lower hemoglobin and leukocyte count, and have a better prognosis than JAK2 V617F and MPL W515 mutated cases. 13 Many techniques have been used to detect CALR mutations, but not all are suitable for a clinical laboratory. Recently, two studies were published comparing different methods of CALR screening. Park et al 27 compared the sensitivity and specificity of HRM, product sizing analysis, and Sanger sequencing. The authors showed that sensitivity was higher for the HRM (96.4%) and product sizing analysis (98.2%), and specificity was 100% for product sizing analysis and Sanger sequencing. According to this study, HRM and product sizing analysis showed similar results, but HRM had a slightly higher rate of false-negative results. In addition, product sizing analysis provided a quantitative measurement of CALR, which could be used to monitor specific treatments. Jones et al 28 compared the limit of detection of Sanger sequencing (10%- 25%), fragment analysis PCR (5%-10%), HRM (5%), and targeted NGS (1.25%). This study showed that each assay was able to detect CALR mutations, but targeted NGS had the best LOD, detecting all mutations to a level of 1%. Conclusion This study shows that diagnosis of patients with clonal and nonclonal hematologic diseases requires the integration of clinical history and cytogenetics or molecular genetic tests (or both when appropriate). The clinical utility of both known and novel molecular biomarkers has substantially increased in recent years. The rapid detection of these biomarkers in patient samples has become a critical component of precision medicine, providing not only valuable diagnostic and prognostic information but also potential treatment targets. Here we describe three independent methods using different technological platforms to detect CALR frameshift mutations in exon 9 that can be easily integrated into routine clinical laboratories. Although the CALR fragment analysis PCR and Sanger sequencing assays can detect only known biomarkers, the NGS test allows for the rapid detection of both known and novel biomarkers. Our institution implemented an algorithm whereby all suspected MPN cases are screened not only for JAK2 (exons 12 and 14), MPL,andCALR mutations but also for 51 genes using the targeted NGS myeloid panel validated in our laboratory. This testing process provides the oncologists with additional information to confirm the final diagnosis and to help define potential treatment strategies for patient management. Corresponding author: Gregory J. Tsongalis, PhD, Dept of Pathology and Laboratory Medicine, DHMC, 1 Medical Center Dr, Lebanon, NH 03756; gregory.j.tsongalis@hitchcock.org. References 1. Harrison CN, Bareford D, Butt N, et al. Guideline for investigation and management of adults and children presenting with a thrombocytosis. Br J Haematol. 2010;149:352-375. 2. McMullin MF, Reilly JT, Campbell P, et al. Amendment to the guideline for diagnosis and investigation of polycythaemia/erythrocytosis. Br J Haematol. 2007;138:821-822. 3. Swerdlow SH, Campo E, Harris NL, et al. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. 4th ed. Lyon, France: IARC Press; 2008. 4. Barbui T, Barosi G, Birgegard G, et al. Philadelphia-negative classical myeloproliferative neoplasms: critical concepts and management recommendations from European LeukemiaNet. J Clin Oncol. 2011;29:761-770. 5. Scott LM, Tong W, Levine RL, et al. JAK2 exon 12 mutations in polycythemia vera and idiopathic erythrocytosis. N Engl J Med. 2007;356:459-468. 6. Butcher CM, Hahn U, To LB, et al. Two novel JAK2 exon 12 mutations in JAK2V617F-negative polycythaemia vera patients. Leukemia. 2008;22:870-873. 7. Campbell PJ, Green AR. The myeloproliferative disorders. N Engl J Med. 2006;355:2452-2466. 8. Pikman Y, Lee BH, Mercher T, et al. MPLW515L is a novel somatic activating mutation in myelofibrosis with myeloid metaplasia. PLoS Med. 2006;3:e270. 9. Rumi E, Pietra D, Guglielmelli P, et al. Acquired copyneutral loss of heterozygosity of chromosome 1p as a molecular event associated with marrow fibrosis in MPLmutated myeloproliferative neoplasms. Blood. 2013;121:4388-4395. 10. Klampfl T, Gisslinger H, Harutyunyan AS, et al. Somatic mutations of calreticulin in myeloproliferative neoplasms. N Engl J Med. 2013;369:2379-2390. 11. Rumi E, Pietra D, Ferretti V, et al. JAK2 or CALR mutation status defines subtypes of essential thrombocythemia with substantially different clinical course and outcomes. Blood. 2014;123:1544-1551. 12. Andrikovics H, Krahling T, Balassa K, et al. Distinct clinical characteristics of myeloproliferative neoplasms with calreticulin mutations. Haematologica. 2014;99:1184-1190. 13. Tefferi A, Lasho TL, Finke CM, et al. CALR vs JAK2 vs MPL-mutated or triple-negative myelofibrosis: clinical, cytogenetic and molecular comparisons. Leukemia. 2014;28:1472-1477. 14. Michalak M, Groenendyk J, Szabo E, et al. Calreticulin, a multi-process calcium-buffering chaperone of the endoplasmic reticulum. Biochem J. 2009;417:651-666. 15. Wang W-A, Groenendyk J, Michalak M. Calreticulin signaling in health and disease. Int J Biochem Cell Biol. 2012;44:842-846. 454 Am J Clin Pathol 2016;146:448-455 American Society for Clinical Pathology 454

16. Gold LI, Eggleton P, Sweetwyne MT, et al. Calreticulin: non-endoplasmic reticulum functions in physiology and disease. FASEB J. 2010;24:665-683. 17. Chao MP, Majeti R, Weissman IL. Programmed cell removal: a new obstacle in the road to developing cancer. Nat Rev Cancer. 2012;12:58-67. 18. Krysko DV, Garg AD, Kaczmarek A, et al. Immunogenic cell death and DAMPs in cancer therapy. Nat Rev Cancer. 2012;12:860-875. 19. Luo B, Lee AS. The critical roles of endoplasmic reticulum chaperones and unfolded protein response in tumorigenesis and anticancer therapies. Oncogene. 2013;32:805-818. 20. Ha JS, Kim YK. Calreticulin exon 9 mutations in myeloproliferative neoplasms. Ann Lab Med. 2015;35:22-27. 21. Bilbao-Sieyro C, Santana G, Moreno M, et al. High resolution melting analysis: a rapid and accurate method to detect CALR mutations. PLoS One. 2014;9:e103511. 22. Chi J, Manoloukos M, Pierides C, et al. Calreticulin mutations in myeloproliferative neoplasms and new methodology for their detection and monitoring. Ann Hematol. 2014;94:399-408. 23. Lim K, Lin H, Chen CG, et al. Rapid and sensitive detection of CALR exon 9 mutations using high-resolution melting analysis. Clin Chim Acta. 2014;440:133-139. 24. Maier CL, Fisher KE, Jones HH, et al. Development and validation of CALR mutation testing for clinical diagnosis. Am J Clin Pathol. 2015;144:738-745. 25. Mehrotra M, Luthra R, Singh RR, et al. Clinical validation of a multipurpose assay for detection and genotyping of CALR mutations in myeloproliferative neoplasms. Am J Clin Pathol. 2015;144:746-755. 26. Gangat N, Wassie E, Lasho T, et al. Mutations and thrombosis in essential thrombocythemia: prognostic interaction with age and thrombosis history. Eur J Haematol. 2015;94:31-36. 27. Park JH, Sevin M, Ramla S, et al. Calreticulin mutations in myeloproliferative neoplasms: comparison of three diagnostic methods. PLoS One. 2015;10:e0141010. 28. Jones AV, Ward D, Lyon M, et al. Evaluation of methods to detect CALR mutations in myeloproliferative neoplasms. Leuk Res. 2015;39:82-87. American Society for Clinical Pathology Am J Clin Pathol 2016;146:448-455 455 455