Review Article Pan-TRK Immunohistochemistry An Example-Based Practical Approach to Efficiently Identify Patients With NTRK Fusion Cancer Esther Conde, MD, PhD; Susana Hernandez, PhD; Elena Sanchez; Rita Maria Regojo, MD; Carmen Camacho, MD; Marta Alonso; Rebeca Martinez; Fernando Lopez-Rios, MD, PhD  Context.—Food and Drug Administration–approved TRK inhibitors with impressive overall response rates are now available for patients with multiple cancer types that harbor NTRK rearrangements, yet the identification of NTRK fusions remains a difficult challenge. These alter- ations are highly recurrent in extremely rare malignancies or can be detected in exceedingly small subsets of common tumor types. A 2-step approach has been proposed, involving a screening by immunohistochemistry (IHC) followed by a confirmatory method (fluorescence in situ hybridization, reverse transcriptase–polymerase chain re- action, or next-generation sequencing) in cases expressing the protein. However, there is no interpretation guide for any of the available IHC clones. Objective.—To provide a pragmatic update on the use of pan-TRK IHC. Selected examples of the different IHC staining patterns across multiple histologies are shown. Data Sources.—Primary literature review with PubMed, combined with personal diagnostic and research experi- ence. Conclusions.—In-depth knowledge of pan-TRK IHC will help pathologists implement a rational approach to the detection of NTRK fusions in human malignancies. (Arch Pathol Lab Med. 2021;145:1031–1040; doi: 10.5858/arpa.2020-0400-RA) S everal TRK inhibitors with impressive overall response rates in patients with NTRK rearrangements are currently available or under clinical development. 1,2 The search for NTRK fusions should benefit from what we have learned in recent years about identifying other druggable rearrangements, mainly in lung cancer (ALK, ROS1, etc). 3–7 Therefore, NTRK fusions can be detected with immuno- histochemistry (IHC), fluorescence in situ hybridization (FISH), reverse transcriptase–polymerase chain reaction, or next-generation sequencing (NGS). 8 If NGS is not routinely performed in all advanced malignant tumors, most proposed algorithms use IHC as a screening method, followed by orthogonal confirmation of all positive IHC cases (mainly using FISH or NGS). 8–11 To further enrich for NTRK fusions, both histology-based and genomic-based triaging approaches have been proposed. 12 A summary of the available evidence is presented in the Table. Until NGS becomes the main testing methodology on all advanced cancers, algorithm considerations should include feasibil- ity, cost, sample size, and pretest probability of NTRK fusions. A review of the frequencies of NTRK fusions highlights the need to be aware of these strategies. NTRK fusions have been observed in 0.31% of adult tumors and 0.34% of pediatric tumors. 13 In clinical series, the most common partners have been NTRK1 and NTRK3. 14–16 The most frequent fusion is ETV6-NTRK3. 16 With the exception of gliomas, 13,15 NTRK2 fusions appear to be restricted to isolated examples of sarcomas or lung adenocarcino- mas. 15,17–19 Although it may be too soon to draw definitive conclusions on the positive and negative predictive value of pan-TRK IHC by partner and cancer type, the largest series to date showed an overall sensitivity of 87.9% and specificity of 81.1%. 14 Decreased sensitivity was reported for NTRK3 fusions and sarcomas, and lower specificity Accepted for publication August 20, 2020. Published online October 28, 2020. From Pathology and Laboratory of Therapeutic Targets, Hospital Universitario HM Sanchinarro, HMHospitales, CIBERONC, Madrid, Spain (Conde, Lopez-Rios); Pathology and Laboratory of Therapeutic Targets, Hospital Universitario HM Sanchinarro, HMHospitales, Madrid, Spain (Hernandez, Sanchez, Alonso, Martinez); Pathology, Hospital Universitario La Paz, Madrid, Spain (Regojo); and Pathol- ogy, Complejo Hospitalario Universitario Insular Materno-Infantil, Las Palmas de Gran Canaria, Spain (Camacho). Conde and Hernandez contributed equally as co–first authors. This study was mainly funded by Roche Spain. We also thank Instituto de Salud Carlos III (ISCIII) (Fondos FEDER and Plan Estatal I þ Dþ I 2008–2011 [PI11/02866] and 2013–2016 [PI14-01176 y PI17- 01001]) and the iLUNG Program (B2017/BMD-3884) from the Comunidad de Madrid. Thermo Fisher Scientific provided an unrestricted grant for the optimization of the next-generation sequencing. Conde receives honoraria from Roche and Pfizer; travel expenses were covered by Roche, Pfizer, and Merck Sharp & Dohme. Lopez- Rios received honoraria from Lilly, Roche, Thermo Fischer Scientific, Pfizer, Bristol-Myers Squibb, Bayer, AstraZeneca, Merck Sharp & Dohme; research funding was received from Lilly, Roche, and Thermo Fischer Scientific. The other authors have no relevant financial interest in the products or companies described in this article. Corresponding author: Fernando Lopez-Rios, MD, PhD, Patholo- gy-Laboratory of Therapeutic Targets, Hospital Universitario HM Sanchinarro, HMHospitales, C/O˜ na, 10. 28050 Madrid, Spain (email: flopezrios@hmhospitales.com). Arch Pathol Lab Med—Vol 145, August 2021 Pan-TRK Immunohistochemistry—Conde et al 1031