The Use of Superlatives in Cancer Research

Letters

from work, child care, and transportation could also improve 4. Unger JM, Coltman CA Jr, Crowley JJ, et al. Impact of the year 2000 access to clinical trials for lower-income patients. Future Medicare policy change on older patient enrollment to cancer clinical trials. J Clin Oncol. 2006;24(1):141-144. research should investigate how to overcome financial barri- 5. Zafar SY, Peppercorn JM, Schrag D, et al. The financial toxicity of cancer ers to clinical trial participation. treatment: a pilot study assessing out-of-pocket expenses and the insured The identification of patient income level as an indepen- cancer patient’s experience. Oncologist. 2013;18(4):381-390. dent predictor of trial participation is important for multiple 6. Dickert N, Grady C. What’s the price of a research subject? approaches to reasons. If income is associated with health status, then im- payment for research participation. N Engl J Med. 1999;341(3):198-203. proving representativeness of lower-income patients in trials would improve the generalizability of study outcomes. Also, The Use of Superlatives in Cancer Research greater participation of lower-income patients would allow The language used in oncology practice and research may elicit trials to be conducted more quickly, speeding the develop- important connotations.1 Whereas most new cancer drugs af- ment of new treatments. Crucially, since clinical trial treat- ford modest benefits,2 approved drugs or those in develop- ments represent the newest available treatments, access to this ment may be heralded as “game changers” or “break- vital resource should be available to individuals of all income throughs” in the lay press. These news articles may be important levels. sources of information to patients, the public, and investors— with a broader reach than medical journal articles. However, Joseph M. Unger, PhD omission of medical context or use of inflated descriptors may 3 Julie R. Gralow, MD lead to misunderstandings among readers. Kathy S. Albain, MD We sought to investigate the use of modest and superla- Scott D. Ramsey, MD tive descriptors in contemporary news articles regarding can- Dawn L. Hershman, MD cer drugs. We sought to determine who uses this inflated language and what classes of drugs were most heralded.

Author Affiliations: SWOG Statistical Center, Fred Hutchinson Cancer Research Center, Seattle, Washington (Unger); Seattle Cancer Care Alliance, University of Methods | We searched 10 superlative terms in conjunction with Washington, Seattle (Gralow); Loyola University, Chicago Stritch School of “cancer drug” in Google’s news search (http://news.google Medicine, Maywood, Illinois (Albain); University of Washington Medical Center, .com) between June 21, 2015, and June 25, 2015. Superlative Seattle (Ramsey); Fred Hutchinson Cancer Research Center, Seattle, Washington (Ramsey); Columbia University, New York, terms included “breakthrough,” “game changer,” “miracle,” New York (Hershman). “cure,” “home run,” “revolutionary,” “transformative,” “life Corresponding Author: Joseph M. Unger, PhD, SWOG Statistical Center, saver,”“groundbreaking,”and “marvel.”Terms were prespeci- Fred Hutchinson Cancer Research Center, 1100 Fairview Ave N, M3-C102, fied and identified through discussion. PO Box 19024, Seattle, WA 98109-1024 ([email protected]). All articles resulted were read in full by one reviewer Published Online: October 15, 2015. doi:10.1001/jamaoncol.2015.3924. (M.V.A.). The following information was extracted: drug de- Author Contributions: Dr Unger had full access to all of the data in the study scribed, mechanism of action, class of medication, whether the and takes responsibility for the integrity of the data and the accuracy of the data agent described had already received US Food and Drug Ad- analysis. Study concept and design: Unger, Ramsey, Hershman. ministration approval, whether the data described con- Acquisition, analysis, or interpretation of data: Unger, Gralow, Albain, Ramsey, cerned human trial results or preclinical (eg, mouse or cell cul- Hershman. ture) data, and the quoted individual (physician, journalist, Drafting of the manuscript: Unger, Hershman. industry expert, or patient). An academic hematologist- Critical revision of the manuscript for important intellectual content: Unger, Gralow, Albain, Ramsey, Hershman. oncologist (V.P.) researched mechanism of action of all drugs Statistical analysis: Unger. and coded their class as cytotoxic, targeted, immunotherapy— Obtained funding: Gralow, Ramsey. checkpoint inhibitor, immunotherapy—therapeutic vaccine, ra- Administrative, technical, or material support: Gralow. Study supervision: Ramsey, Hershman. diotherapy, gene therapy, or other. Conflict of Interest Disclosures: None reported. Results | We found 94 news articles from 66 distinct news out- Funding/Support: This work was supported by a Breast Cancer Research Foundation grant and by the National Institutes of Health, National Cancer lets that made 97 superlative mentions that fit our criteria, re- Institute (NCI) Community Oncology Research Program Research Base grant ferring to 36 specific drugs, with 3 articles not naming the drug. 5UG1CA189974-01. The most common class of drugs referenced was targeted therapy Role of the Funder/Sponsor: The funding organizations had a role in the design (17 of 36 [47%]). Nine (25%) cytotoxic drugs were discussed, fol- and conduct of the study and in the collection and management of the data but lowed by 5 (14%) immunotherapy checkpoint inhibitors, 3 (8%) had no role in the analysis and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit for publication. cancer vaccines, 1 radiotherapy, and 1 gene therapy. 1. Murthy VH, Krumholz HM, Gross CP. Participation in cancer clinical trials: Among 97 superlatives used, 39 (40%) referred to a tar- race-, sex-, and age-based disparities. JAMA. 2004;291(22): geted therapy, 37 (38%) referred to an immunologic check- 2720-2726. point inhibitor, 10 (10%) referenced a cytotoxic drug, 5 (5%) 2. Unger JM, Hershman DL, Albain KS, et al. Patient income level and cancer discussed a therapeutic cancer vaccine, 3 (3%) did not name clinical trial participation. J Clin Oncol. 2013;31(5):536-542. the drug, 2 (2%) referred to a radiotherapy, and 1 (1%) referred 3. Javid SH, Unger JM, Gralow JR, et al. A prospective analysis of the influence to gene therapy. of older age on physician and patient decision-making when considering enrollment in breast cancer clinical trials (SWOG S0316). Oncologist. 2012;17(9): Precisely half (18 of 36) of drugs described had not re- 1180-1190. ceived Food and Drug Administration approval (as of July 15,

jamaoncology.com (Reprinted) JAMA Oncology January 2016 Volume 2, Number 1 139

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Letters

Table. Frequency and Characteristics of Cancer Drugs Described With Superlatives

Superlative Frequency, No. (%)a Superlative(s) Drug FDA-Approved Clinical Drug (N = 97) Used (Frequency) Classification Drug(s) Data? Ipilimumab and 20 (21) Breakthrough (7), Immunotherapy— Yes Yes nivolumab miracle (5), game checkpoint inhibitor (Yervoy-Opdivo changer (5), combination) revolutionary (2), groundbreaking (1) Pembrolizumab 12 (12) Revolutionary (5), Immunotherapy— Yes Yes (Keytruda) game changer (2), checkpoint inhibitor groundbreaking (2), cure (2), miracle (1) Palbociclib (Ibrance) 10 (10) Groundbreaking (6), Targeted therapy Yes Yes game changer (2), revolutionary (1), miracle (1) Trastuzumab 7 (7) Revolutionary (4), Targeted therapy Yes Yes emtansine (Kadcyla) miracle (3) Dinutuximab 4 (4) Game changer (1), Targeted therapy Yes Yes (Unituxin) groundbreaking (1), Abbreviation: FDA, Food and Drug breakthrough (1), Administration. miracle (1) a MPDL3280A 3 (3) Game changer (2), Immunotherapy— No Yes Drugs with 1 (1%) superlative: revolutionary (1) checkpoint inhibitor ABT-199, acelarin, ALM201, Olaparib (Lynparza) 3 (3) Revolutionary (2), Targeted therapy Yes Yes bortezomib (Velcade), brentuximab breakthrough (1) vedotin (Adcetris), Cimavax, T-VEC 3 (3) Breakthrough (3) Immunotherapy— No Yes docetaxel (Taxotere), doxorubicin vaccine with vinorelbine, epidaza (Chidamide), eribulin (Halaven), Pertuzumab (Perjeta) 3 (3) Groundbreaking (3) Targeted therapy Yes Yes figitumumab, G1 T-28, ibrutinib Unnamed 3 (3) Breakthrough (1), ………(Imbruvica), ipilimumab, NeuVax, miracle (1), nivolumab (Opdivo), OH14 game changer (1) compound, OTS964, PENAO, Radium-223 2 (2) Game changer (2) Radiotherapeutic drug Yes Yes protein-bound paclitaxel dichloride (Alpharadin or Xofigo) (Abraxane), rociletinib, TargomiRs, TRXE-009, vemurafenib (Zelboraf), BPM31510 2 (2) Revolutionary (2) Cytotoxic therapy No Yes ZL105.

2015) for at least 1 indication. For 5 of 36 (14%) drugs, super- cology conference; as such, it may not reflect the use of these latives were used in the absence of clinical data (ie, based solely terms at other times, as well as in other years. on mouse, cell culture, and/or preclinical work). The specific A range of speakers used superlatives, but the majority drug mentioned, superlative used, and other characteristics were journalists (55%), who may not have the expertise to iden- are described in the Table. tify the most promising medical therapies, or what magni- A variety of speakers were credited with using the super- tude of benefit warrants a superlative. The use of superla- lative (53 journalists [55%], 26 physicians [27%], 9 industry ex- tives is common in cancer research news articles. Some of this perts [9%], 8 patients [8%], and 1 member of US Congress [1%]). use may be questioned. In the majority of cases (55%) the superlative was used by the author of the article without any other attribution. Matthew V. Abola, BA Vinay Prasad, MD, MPH Discussion | Our investigation finds that the use of superlatives to describe approved (50%) and nonapproved cancer drugs Author Affiliations: Medical student at Case Western Reserve University School of Medicine, Cleveland, Ohio (Abola); Medical Oncology Service, (50%) is common. Superlatives are used for all types of medi- National Cancer Institute, National Institutes of Health, Bethesda, cations, including those, such as therapeutic cancer vac- Maryland (Prasad); now with Division of Hematology and Medical Oncology, cines, which historically have low response rates4 and drugs Knight Cancer Institute, Oregon Health and Science University, Portland (Prasad). that have not yet shown overall survival benefits (eg, palbociclib). Of concern, 14% of drugs were praised without any hu- Corresponding Author: Vinay Prasad, MD, MPH, Division of Hematology and Medical Oncology, Knight Cancer Institute, Oregon Health and Science man data. University, 3181 SW Sam Jackson Park Rd, Portland, OR 97239 Use of superlatives reflects the current “hot fields” of can- ([email protected]). cer research. Although immunologic checkpoint inhibitors Published Online: October 29, 2015. doi:10.1001/jamaoncol.2015.3931. made up only 14% of the unique drugs mentioned, they ac- Author Contributions: Mr Abola had full access to all of the data in the study counted for 38% of all superlatives. Targeted therapies were and takes responsibility for the integrity of the data and the accuracy of the data the most common drugs mentioned, and they received the analysis. Study concept and design: Both authors. most superlatives. One limitation to our work is that it was con- Acquisition, analysis, or interpretation of data: Both authors. ducted 3 weeks after the 2015 American Society of Clinical On- Drafting of the manuscript: Both authors.

140 JAMA Oncology January 2016 Volume 2, Number 1 (Reprinted) jamaoncology.com

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Letters

Critical revision of the manuscript for important intellectual content: Abola. (van Gool, Bosse); Molecular and Population Genetics Laboratory, Wellcome Statistical analysis: Abola. Trust Centre for Human Genetics, University of Oxford, Oxford, England Administrative, technical, or material support: Abola. (Church). Study supervision: Prasad. Corresponding Author: David N. Church, MBChB, DPhil, Molecular and Conflict of Interest Disclosures: None reported. Population Genetics Laboratory, Wellcome Trust Centre for Human Genetics, 1. Ellis LM, Blanke CD, Roach N. Losing “losing the battle with cancer.” JAMA Oncol. University of Oxford, Roosevelt Dr, Oxford OX3 7BN, England 2015;1(1):13-14. ([email protected]). 2. Fojo T, Mailankody S, Lo A. Unintended consequences of expensive cancer Conflict of Interest Disclosures: None reported. therapeutics—the pursuit of marginal indications and a me-too mentality that 1. Howitt BE, Shukla SA, Sholl LM, et al. Association of polymerase e–mutated stifles innovation and creativity: the John Conley Lecture. JAMA Otolaryngol and microsatellite-instable endometrial cancers with neoantigen load, number Head Neck Surg. 2014;140(12):1225-1236. of tumor-infiltrating lymphocytes, and expression of PD-1 and PD-L1. JAMA Oncol. 3. Dentzer S. Communicating medical news—pitfalls of health care journalism. 2015;1(9):1319-1323. N Engl J Med. 2009;360(1):1-3. 2. Church DN, Briggs SE, Palles C, et al; NSECG Collaborators. DNA polymerase ε and δ exonuclease domain mutations in endometrial cancer. Hum Mol Genet. 4. Horstmann E, McCabe MS, Grochow L, et al. Risks and benefits of phase 1 2013;22(14):2820-2828. oncology trials, 1991 through 2002. N Engl J Med. 2005;352(9):895-904. 3. Kandoth C, Schultz N, Cherniack AD, et al; Cancer Genome Atlas Research Network. Integrated genomic characterization of endometrial carcinoma. Nature. 2013;497(7447):67-73. COMMENT & RESPONSE 4. Church DN, Stelloo E, Nout RA, et al. Prognostic significance of POLE proofreading mutations in endometrial cancer. J Natl Cancer Inst. 2015;107(1): Neoepitopes and CD3-Positive 402-410. and CD8-Positive Cells in Polymerase e–Mutated 5. van Gool IC, Eggink FA, Freeman-Mills L, et al. POLE proofreading mutations and Microsatellite-Instable Endometrial Cancers elicit an antitumor immune response in endometrial cancer. Clin Cancer Res. To the Editor We read with interest the Brief Report by Howitt 2015;21(14):3347-3355. and colleagues.1 Over the last few years, we have also studied the interesting subgroup of endometrial cancers (ECs) with mu- In Reply Our study1 was initially submitted in January 2015 as tation in the POLE proofreading exonuclease domain,2,3 and an abstract for the 2015 American Society of Clinical Oncol- in particular their association with excellent prognosis,3,4 not ogy Annual Meeting and was accepted in March 2015 as an oral mentioned by Howitt et al1 in their report. presentation; it was subsequently submitted to JAMA Oncol- In their study,Howitt et al1 claim that their analysis is the first ogy in April 2015 before the study by van Gool et al2 appeared to demonstrate increased predicted neoepitopes and numbers online. The goal of our study was to address whether hyper- of CD3-positive and CD8-positive cells in POLE-mutant and mutated (microsatellite-unstable [MSI]/POLE-mutant) endo- microsatellite-unstable (MSI) ECs. However, in a study published metrial cancers (ECs) were more immunogenic compared with earlier this year,5 we showed that both POLE proofreading- microsatellite-stable (MSS) ECs, and not whether this associa- mutant and MSI ECs demonstrate significantly greater CD8- tion may account for the improved survival of POLE-mutated positive cell infiltration and higher numbers of predicted neo- tumors; accordingly, no reference was made to studies report- antigens than microsatellite-stable ECs. In this analysis, we also ing an association of POLE-mutated tumors with improved demonstrated that POLE-mutant and, to a lesser extent, MSI ECs survival. display increased expression of genes encoding immune check- We would like to highlight 3 important differences be- point molecules, including PDCD1 and PDL1. Furthermore, as our tween the study by van Gool et al2 and ours. First, our study study included a substantially greater number of POLE-mutant also demonstrated that PD-1 expression is significantly in- tumors than that of Howitt et al1 (47 compared with 3 cases), we creased in tumor-associated lymphocytes of MSI/POLE- were also able to confirm the postulate of Howitt et al that the mutated ECs. While van Gool et al2 briefly mentioned the ra- greater number of neoantigens in POLE-mutant than in MSI ECs tionale for immunotherapy with PD-1 and PD-L1 inhibitors, they is reflected in significantly greater CD8-positive cell infiltration.5 evaluated PD-1 expression by RNaseq on EC samples from the Collectively, the small study by Howitt et al,1 and our ear- Cancer Genome Atlas data set and inferred it on the basis of lier, more comprehensive analysis, suggest that both POLE CD8A expression. In our study, we directly visualized the in- proofreading-mutant and MSI tumors are more immuno- creased expression of PD-1 and CD8 by using immunohisto- genic than other ECs. We therefore agree with Howitt et al1 that chemical analysis on serial sections. Interestingly, van Gool et these cancers may be excellent candidates for immune check- al2 did not find a significant difference in PD-1 expression be- point inhibitor therapy, as indeed we have previously tween MSI and MSS tumors, which is in contrast to our find- suggested.5 We also suggest that the striking immune re- ings using immunohistochemical analysis. sponse observed in POLE-proofreading mutant ECs may con- Second, we performed immunohistochemical analysis for tribute to their favorable clinical outcome. PD-L1 expression and also described increased PD-L1 expression in intraepithelial immune cells of MSI/POLE-mutated EC Inge C. van Gool, MD compared with MSS EC but did not find significant immuno- Tjalling Bosse, MD, PhD histochemical expression of PD-L1 in tumor cells (save for 1 David N. Church, MBChB, DPhil POLE-mutated EC with strong diffuse membranous expression). Of note, response to anti–PD-L1 therapy has been shown Author Affiliations: Department of Pathology, Leiden University Medical to correlate with expression of PD-L1 in tumor-infiltrating im- Center, Albinusdreef 2, Postbus 9600, 2300 RC Leiden, the Netherlands mune cells but not in tumor cells.3 We suggest that immuno-

jamaoncology.com (Reprinted) JAMA Oncology January 2016 Volume 2, Number 1 141

Downloaded From: https://jamanetwork.com/ on 10/02/2021