DOI: 10.1200/JCO.19.00318 Journal of Clinical Oncology - published online before print May 22, 2019
Analysis of the Association Between Adverse Events and Outcome in Patients Receiving a Programmed Death Protein 1 or Programmed Death Ligand 1 Antibody
J.A.B. and R.P. contributed equally to this work and are co-senior authors.
To assess the relationship among tumor response rate, overall survival, and the development of related adverse events of special interest (AESIs) or related immune-mediated adverse events (imAEs) in patients with urothelial cancer treated with anti–programmed death protein 1 or ligand 1 (anti–PD-1/L1) antibodies.
We examined seven trials in 1,747 patients with metastatic or locally advanced urothelial cancer that led to approval of an anti–PD-1/L1 antibody. Five trials enrolled patients who had received prior platinum-based therapy, and two enrolled patients who were cisplatin ineligible. The data sets were searched for AESIs, related AESIs, imAEs, and related imAEs. The relationship to study drug was determined by the investigator. ImAEs were defined as AESIs treated with topical or systemic corticosteroids.
In these exploratory analyses, a related AESI was reported in 64% of responding patients and in 34% of patients who did not respond to the anti–PD-1/L1 antibody, whereas a related imAE occurred in 28% and 12% of patients who did and did not respond to study drug, respectively. In a responder analysis, an increase in overall survival was seen in patients with related AESIs compared with those with no related AESIs (hazard ratio, 0.45; 95% CI, 0.39 to 0.52). Fifty-seven percent of responding patients with a related AESI reported the AESI before documentation of response.
Patients who responded to treatment with an anti–PD-1/L1 antibody were more likely to report a related AESI or related imAE. This relationship did not seem to be due to the increased duration of exposure in responding patients. Systemic corticosteroid use did not appear to affect the duration of response.
Five monoclonal antibodies directed against programmed death protein 1 (PD-1) or programmed death protein ligand 1 (PD-L1) have been approved in the United States for the treatment of advanced or metastatic urothelial cancer. All have been approved for the treatment of patients who received prior platinum-based therapy, and two were also approved for the treatment of patients who are cisplatin ineligible.
A relationship between patient outcome and the development of characteristic adverse events (AEs) has been seen with nonimmunotherapy agents. For example, the development of rash after treatment with epidermal growth factor receptor inhibitors was initially associated with patient outcome.1,2 Given the immune mechanism of action of anti–PD-1/L1 antibodies, it is reasonable to associate the development of autoimmune events with improved outcome because activation of the immune system could lead to both tumor response and autoimmunity. It is unclear whether the development of autoimmune AEs in patients with melanoma who receive ipilimumab affects outcome, although the development of autoimmune AEs has been associated with improved outcome with nivolumab.3-5
Data from the trials that have led to the approvals of five anti–PD-1/L1 antibodies in the treatment of urothelial cancer provide an opportunity to study the interplay between patient outcome and the development of AEs of special interest (AESIs) and immune-mediated AEs (imAEs) that are related to the use of study drug. Because of concerns about the effect of corticosteroid use on patient outcome, imAEs were examined separately. To investigate the relationship between patient outcome and related AESIs/imAEs, we performed a pooled analysis of these trials that permitted an examination of the association between related AESIs/imAEs and tumor response, the timing of response, and overall survival (OS) as well as the effect of corticosteroid use on the duration of response.
Trials were eligible for this pooled analysis if they were submitted to the US Food and Drug Administration to support a marketing application that was ultimately approved for the treatment of patients with advanced or metastatic urothelial cancer. Individual patient data were pooled.
The study designs of the included trials were similar, and the results have been published elsewhere (Table 1).6-12 All trials enrolled patients with locally advanced or metastatic urothelial cancer. Five trials enrolled patients who had received prior platinum-based therapy (cisplatin or carboplatin), and two enrolled patients who were cisplatin ineligible. Cisplatin ineligible was defined as one of the following: creatinine clearance of approximately 30 to 59 mL/min, performance status (PS) of 2, grade 2 or greater peripheral neuropathy, or grade 2 or greater hearing loss. One trial also included class 3 heart failure in the definition of cisplatin ineligible. Note that cisplatin-ineligible patients may have received prior cisplatin but were cisplatin ineligible at the time of enrollment. All trials excluded patients with autoimmune disease, although most permitted the entry of patients with vitiligo, psoriasis, thyroid disease, adrenal disease, or controlled diabetes. Patients who required a corticosteroid dose greater than physiologic replacement and those who required other immunosuppressive medications were excluded from entry. Patients received study drug every 2 to 3 weeks, and tumor assessments occurred every 6 to 9 weeks. In general, treatment continued until radiographic (often requiring confirmation) or clinical progression. In six single-arm trials, the primary end point was response rate. In the seventh randomized trial, response rate was a secondary end point. In all trials, response rate was determined by independent review using Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. The protocols for each study included a list of AESIs. These lists not only focused on autoimmune events but also included broad terms, such as diarrhea or rash, that could occur as a result of other causes.
All analyses should be considered exploratory. The AESIs identified in each protocol were grouped and used to identify AEs in each trial. Because all the studies included at least 30 days of follow-up after the last dose of study drug, the cutoff of 30 days was chosen for AESIs in these analyses. ImAEs were defined as any AESI treated with a topical or systemic corticosteroid. The relationship between AESIs, imAEs, and the study drug was determined by the investigator. Concomitant medication data sets were examined for the use of corticosteroids, immunosuppressive medications, or thyroid medications. Many patients received corticosteroids for indications that did not seem to be immune related (eg, fatigue, nausea) and were not included in the list of AESIs.
Analyses included a descriptive exploratory evaluation of the association between the development of an AESI/imAE or related AESI/imAE and tumor response (complete or partial response). The associations between the development of a related AESI and response or a related imAE and response in patients who used topical corticosteroids or thyroid medications also were examined in those who did not receive systemic corticosteroids during the treatment period.
The relationship between the development of a related AESI/imAE and time in the study was examined by analyzing the time to first related AESI or imAE in patients who did and did not respond to study drug using the stratified Cox regression model, stratified by study. The responder status (responded yes/no) and the exposure duration were included in the model as covariates. Exposure duration was the difference in days between the treatment stop and start dates. An interaction term was included to determine whether there were differences by response and treatment exposure and retained in the model if found to be statistically significant. The association between responder status and development of a related AESI/imAE also was explored using a binary outcome multivariable-stratified logistic regression model, stratified by study. To examine the effect of the duration of exposure on the development of a related AESI/imAE, the responder status, the duration of exposure, and a multiplicative interaction term were included as covariates in the model.
Analyses of the association between related AESIs/imAEs and OS were adjusted for age, sex, PS, and the presence or absence of liver metastases at baseline as covariates in the Cox proportional hazards regression model, with the individual studies as stratification factors to account for the heterogeneity within the studies. Among responders, the duration of response in patients who did or did not receive corticosteroids was examined. These analyses used Kaplan-Meier curves and the Cox proportional hazards regression model, with the individual studies as stratification factors. All analyses were performed using R statistical software (www.r-project.org).
Most of the trials analyzed patient subgroups on the basis of either the Bellmunt or Bajorin risk scores, which were calculated in patients who received prior platinum-based therapy and first-line treatment, respectively.13,14 In these analyses, both risk scores were calculated for each trial (using the sites of disease determined by independent review) so that these results could be compared easily.
Seven multicenter clinical trials in 1,747 patients were pooled for this analysis (Table 1). Patient demographics and baseline characteristics, those who had received prior platinum-based therapy, and those who were cisplatin ineligible are listed in Table 2. Compared with the five trials that enrolled patients who had received prior platinum-based therapy, the patients who were cisplatin ineligible were slightly older, more likely to have a PS of 2, and more likely to have a reduced creatinine clearance.
A relationship was seen between confirmed response by RECIST version 1.1 and the development of a related AESI/imAE in the pooled data (Table 3). Each individual trial demonstrated a substantial association between the development of a related AESI/imAE and response, with related AESIs occurring in 52% to 73% of responding patients. This association was evaluated through a logistic regression model adjusted for the duration of exposure wherein the odds ratio (OR) of having a related AESI was 5.38 (95% CI, 3.06 to 9.46) for responders compared with nonresponders (Appendix Table A6, online only); for related imAEs, the OR was 3.77 (95% CI, 2.02 to 7.03; Appendix Table A5, online only). The interaction was significant in both models, which indicates that the responders and nonresponders had a differential proclivity for developing a related AESI/imAE given the same duration of exposure. Results from the Cox proportional hazards regression models of the time to first related AESI adjusted for exposure found an HR of 1.30 (95% CI, 1.06 to 1.59; Appendix Fig A1, online only). A similar analysis of related imAEs found an HR of 1.67 (95% CI, 1.24 to 2.27, Appendix Fig A2, online only).
It was believed that the relationship between related AESIs/imAEs and response may be stronger in patients who had not received a systemic corticosteroid. This was further examined in patients who received a topical corticosteroid or thyroid medication to treat a related AESI but who did not receive a systemic corticosteroid during the treatment period. The relationship between response and the development of a related AESI, topical corticosteroid use, or the use of thyroid medication seemed to be similar, regardless of the use of a systemic corticosteroid (Table 3).
Given the differences in patient populations, additional examinations included the association of response and related AESIs/imAEs by sex, age (less than 80 years or 80 years and older), prior platinum-based therapy, or cisplatin ineligibility (Appendix Tables A1 and A2, online only). These examinations did not show a substantial change in the association between response and reports of related AESIs/imAEs. This relationship also was examined in patients who received a PD-1 or a PD-L1 inhibitor; in those who received zero, one, or two prior systemic regimens; and in those who had a PS of 0, 1, or 2. These analyses were limited to patients who received atezolizumab or pembrolizumab because both were administered to patients who had received prior platinum-based therapy and to patients who were cisplatin ineligible (Appendix Tables A3 and A4, online only). All patients who had not received a prior systemic regimen and most of those with a PS of 2 were enrolled in the studies of patients who were cisplatin ineligible. Again, these examinations did not show a substantial change in the association between the development of a related AESI or related imAE and tumor response.
A responder analysis of the relationship between the development of a related AESI or imAE and OS when adjusted for baseline covariates found an improvement in OS among patients who developed a related AESI (HR, 0.45; 95% CI, 0.39 to 0.53) or a related imAE (HR, 0.53; 95% CI, 0.43 to 0.66). The relationship between OS and the development of a related AESI is shown in Figure 1, whereas the relationship between OS and the development of a related imAE is shown in Appendix Figure A3 (online only).
After the relationship between patient outcome and the development of a related AESI/imAE was seen, the timing of response relative to the development of these events was examined (Table 4). Patients with a related AESI/imAE both before and after the date of response were counted more than once in this table. Among the 226 responding patients who reported a related AESI, 57% reported an event before documentation of a response. The most common related AESIs before response were pruritus, rash, and thyroid disease. Most were grade 1 to 2 events, but seven patients (5%) had a grade 3 to 4 related AESI. Among the 100 responding patients who had related imAEs, 27 were prescribed a corticosteroid before response. This included six patients who began a systemic corticosteroid at a median of 44 days before response (range, 61 to 11 days before the day of response).
The use of systemic corticosteroids did not seem to negatively affect the chances of developing a response and did not seem to affect the duration of response (Fig 2). Responding patients who did and did not receive systemic corticosteroids (administered for any reason, including an imAE) seemed to have a similar duration of response (HR, 1.09; 95% CI, 0.70 to 1.69).
These exploratory analyses focused on the association between efficacy and the development of related AESIs/imAEs among 1,747 patients with urothelial cancer who received a PD-1/L1 inhibitor in seven clinical trials that led to product approval in the United States. We observed a substantial relationship between the development of a related AESI/imAE and patient outcome in terms of response and OS. This relationship did not seem to be affected by systemic corticosteroids. Likewise, systemic corticosteroids did not seem to affect the duration of response. Our analyses also examined the relationship between the timing of the onset of related AESIs or the timing of the initiation of corticosteroids with related imAEs and the documentation of response. We note that in many cases, the responding patients developed a related AESI or a related imAE before documentation of tumor response.
A key limitation is that the incidence of related AESIs/imAEs among responders was confounded by the duration of treatment. To adjust for duration of exposure, multivariable analyses were conducted, including an assessment of the OR, corrected for the duration of exposure, for the development of related AESIs/imAEs in responding and nonresponding patients and the time to first related AESI/imAE. The ORs, corrected for the duration of exposure, continued to show an association between response and the development of related AESIs/imAEs. The significant interaction term between response and duration of exposure in the logistic regression model suggests that the OR for responders compared with nonresponders is different for patients with different exposure times. The analyses of time to first related AESI/imAE found that when adjusted for exposure, responding patients were 30% more likely to have a related AESI and 67% more likely to have a related imAE.
These analyses also are limited by their exploratory and retrospective nature and do not account for the fact that related AESI/imAEs are treatment-mediated events that occur after the initiation of therapy. The duration of observation for AESIs/imAEs, given that these are biologic therapies with a prolonged half-life, also may have limited accurate collection of these events. In addition, these analyses are limited by the definitions of AESIs and imAEs and the determination of their relationship to study drug (ie, that AESIs are events of interest that may not be immune mediated). AESIs and imAEs that were believed by the investigator to be related to the use of an anti–PD-1/L1 antibody were used in these analyses to increase the likelihood that these events are, in fact, immune mediated. Another limitation is that although imAEs are defined by the use of corticosteroids, the decision to use corticosteroids may vary among practitioners and among the anti–PD-1/L1 antibodies. The fact that a large database and five different anti–PD-1/L1 antibodies were examined may lessen this concern. Furthermore, these analyses did not distinguish among the dose, duration, or route of the corticosteroid to determine whether these variables may have had an additional effect on patient outcomes. Finally, although each trial showed a substantial relationship between the development of a related AESI/imAE and response, there was a difference in the percentage of responding patients with a related AESI/imAE among the seven trials. Additional work is needed on the related AESIs that occurred before response and other factors, such as AESI grade, that may be predictive of patient outcome.
The interplay between imAEs and markers of efficacy, such as response rate and OS, remains a research question with important clinical implications. First, there is the question of whether the same activation of the immune system that leads to the development of antitumor responses could also simultaneously lead to the development of related AESIs/imAEs. This would have implications for counseling patients who may develop such AEs if these events are confirmed to be correlative biomarkers of efficacy outcome. An additional question with clinical implications is whether the use of systemic corticosteroids to abrogate immune activation also might dampen the desired immune-mediated responses. Our analyses provide reassuring data on the effect of systemic corticosteroids on the duration of response. However, this analysis should be viewed with caution because of the limited, nonuniform follow-up between studies.
Much of what was previously known about the relationship between the development of AESIs/imAEs and efficacy was derived from retrospective analyses of patients with melanoma, for whom ipilimumab has been Food and Drug Administration approved since 2011. Some analyses of ipilimumab-treated patients who developed imAEs seem to indicate that these patients may indeed have an increased response rate compared with those who do not develop imAEs and that the use of corticosteroids does not seem to have a negative effect on response rates or durations.3 A landmark analysis, however, found no relationship between imAEs and time to treatment failure.4 Analysis of the use of nivolumab in patients with melanoma, when adjusted for the number of doses of nivolumab, also seemed to show a trend toward improved response rates in patients with imAEs versus those without imAEs. The use of corticosteroids in patients who received nivolumab similarly did not seem to have a detrimental effect on efficacy outcomes.5
Discipline-specific literature also describes the observed interplay between imAEs and tumor responses. One case-control study in the dermatology literature of 20 patients observed that patients with immune-mediated dermatitis on PD-1 inhibitors tended to have higher overall response rates than matched patients who did not develop dermatitis (65% v 17%), and that patients with dermatitis had a longer 1-year OS rate (90% v 57%).15 A similar effect was seen in an analysis of patients with melanoma treated with a PD-1 inhibitor where patients who developed vitiligo seemed more likely to have an objective response to treatment than those who did not develop vitiligo (71% v 28%). This difference was not statistically significant when corrected for lead-time bias.16 The rheumatology literature describes a single-institution experience in which 35 of 524 consecutive checkpoint inhibitor–treated patients were referred to rheumatology for management of rheumatic imAEs. Among patients with rheumatic imAEs, response rates were significantly higher than in patients without imAEs (85.7% v 35.3%). The proportion of responders was also higher in patients who presented with other imAEs compared with those without imAEs (75.1% v 35.3%).17
To our knowledge, this study of seven trials using five anti–PD-1/L1 antibodies in the treatment of a single cancer analyzes the largest patient population to date (n = 1,747) in which the development of related AESIs/imAEs could be examined. We observed a substantial association between the development of related AESIs/imAEs and patient outcome. This relationship was seen both before and after documented response. Equally important is our observation that the use of systemic corticosteroids did not seem to affect the duration of response.
See accompanying Oncology Grand Rounds on page 2714
Conception and design: V. Ellen Maher, Laura Fernandes, Chana Weinstock, Shenghui Tang, Marc Theoret, Julia A. Beaver, Richard Pazdur
Administrative support: Kirsten B. Goldberg
Collection and assembly of data: V. Ellen Maher, Laura L. Fernandes, Chana Weinstock, Sundeep Agrawal, Michael Brave, Yang-min Ning, Harpreet Singh, Daniel Suzman, James Xu
Data analysis and interpretation: V. Ellen Maher, Laura L. Fernandes, Chana Weinstock, Shenghui Tang, Rajeshwari Sridhara, Amna Ibrahim, Marc Theoret, Julia A. Beaver, Richard Pazdur
Manuscript writing: All authors
Final approval of manuscript: All authors
Accountable for all aspects of the work: All authors
The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/site/ifc.
Other Relationship: Grand Rounds
No other potential conflicts of interest were reported.
1. Publication title: JOURNAL OF CLINICAL ONCOLOGY.
2. Publication no.: 0732-183X.
3. Filing date: October 1, 2019.
4. Issue frequency: 36 times/year; 3 times/month.
5. No. of issues published annually: 36.
6. Annual subscription price: $625.00.
7. Complete mailing address of known office of publication: 2318 Mill Road, Suite 800, Alexandria, VA 22314-4609.
8. Complete mailing address of the headquarters or general business offices of the publisher: 2318 Mill Road, Suite 800, Alexandria, VA 22314-4609.
9. Full names and complete mailing addresses of publisher, editor, and managing editor: Publisher: David Sampson, Publisher, Journal of Clinical Oncology, 2318 Mill Road, Suite 800, Alexandria, VA 22314-4609. Editor: Stephen A. Cannistra, MD, Editor-in-Chief, Journal of Clinical Oncology, Beth Israel Deaconess Medical Ctr., 330 Brookline Ave., Boston, MA 02215. Managing Editor: Kenneth G. Kornfield, Managing Editor, Journal of Clinical Oncology, 2318 Mill Road, Suite 800, Alexandria, VA 22314.
10. Owner: American Society of Clinical Oncology, 2318 Mill Road, Suite 800, Alexandria, VA 22314-4609.
11. Known bondholders, mortgagees, and other security holders owning or holding 1 percent or more of total amount of bonds, mortgages, or other securities: None.
12. Purpose, function, and nonprofit status: Has not changed during preceding 12 months.
13. Publication title: JOURNAL OF CLINICAL ONCOLOGY.
14. Issue date for circulation data: Volume 37, Issue 27 (September 20, 2019).
15. Extent and nature of circulation: Average number of copies each issue during preceding 12 months: (a) Total no. copies (net press run), 23,073. (b) Paid and/or requested circulation: (1) Paid/requested outside-county mail subscriptions stated on Form 3541 (include advertiser’s proof and exchange copies): 16,263; (2) Paid in-county subscriptions stated on Form 3541 (include advertiser’s proof and exchange copies): N/A; (3) Sales through dealers and carriers, street vendors, counter sales, and other non-USPS paid distribution: 5,361; (4) Other classes mailed through the USPS: N/A. (c) Total paid and/or requested circulation (sum of 15b (1), (2), (3), and (4)): 21,624. (d) Free distribution by mail (samples, complimentary, and other free): (1) Outside-county as stated on form 3541: 51; (2) In-county as stated on form 3541: N/A; (3) Other classes mailed through the USPS: N/A; (4) Free distribution outside the mail (carriers or other means): 458. (e) Total free distribution (sum of 15d (1), (2), (3), and (4)): 509. (f) Total distribution (sum of 15c and 15e): 22,133. (g) Copies not distributed: 939. (h) Total (sum of 15f and 15g): 23,073. (i) Percent paid and/or requested circulation (15c/15f x 100): 97.90%. Actual no. copies of single issue published nearest to filing date: (a) Total no. copies (net press run): 23,541. (b) Paid and/or requested circulation: (1) Paid/requested outside-county mail subscriptions stated on Form 3541 (include advertiser’s proof and exchange copies): 17,324; (2) Paid in-county subscriptions stated on Form 3541 (include advertiser’s proof and exchange copies): N/A; (3) Sales through dealers and carriers, street vendors, counter sales, and other non-USPS paid distribution: 5,435; (4) Other classes mailed through the USPS: N/A. (c) Total paid and/or requested circulation (sum of 15b (1), (2), (3), and (4)): 22,759. (d) Free distribution by mail (samples, complimentary, and other free copies): (1) Outside-county as stated on Form 3541: 2; (2) In-county as stated on Form 3541: N/A; (3) Other classes mailed through the USPS: N/A; (4) Free distribution outside the mail (carriers or other means): 523. (e) Total free distribution (sum of 15d (1), (2), (3), and (4)): 525. (f) Total distribution (sum of 15c and 15e): 23,284. (g) Copies not distributed: 257. (h) Total (sum of 15f and 15g): 23,541. (i) Percent paid and/or requested circulation (15c/15f x 100): 97.75%.
16. Total circulation includes electronic copies. Report circulation on PS Form 3526-X worksheet. N/A.
17. This Statement of Ownership will be printed in Volume 37, Issue 30 (October 20, 2019).
18. I certify that the statements made by me above are correct and complete.
David Sampson, Publisher
We thank the patients, staff, and investigators who participated in these seven trials in addition to the pharmaceutical companies who sponsored the trials: AstraZeneca, Bristol-Myers Squibb, EMD Serono, Genentech, and Merck.
|1.||Peréz-Soler R, Saltz L: Cutaneous adverse effects with HER1/EGFR-targeted agents: Is there a silver lining? J Clin Oncol 23:5235-5246, 2005 Link, Google Scholar|
|2.||Abdel-Rahman O, Fouad M: Correlation of cetuximab-induced skin rash and outcomes of solid tumor patients treated with cetuximab: A systematic review and meta-analysis. Crit Rev Oncol Hematol 93:127-135, 2015 Crossref, Google Scholar|
|3.||Beck KE, Blansfield JA, Tran KQ, et al: Enterocolitis in patients with cancer after antibody blockade of cytotoxic T-lymphocyte-associated antigen 4. J Clin Oncol 24:2283-2289, 2006 Link, Google Scholar|
|4.||Horvat TZ, Adel NG, Dang TO, et al: Immune-related adverse events, need for systemic immunosuppression, and effects on survival and time to treatment failure in patients with melanoma treated with ipilimumab at Memorial Sloan Kettering Cancer Center. J Clin Oncol 33:3193-3198, 2015 Link, Google Scholar|
|5.||Weber JS, Hodi FS, Wolchok JD, et al: Safety profile of nivolumab monotherapy: A pooled analysis of patients with advanced melanoma. J Clin Oncol 35:785-792, 2017 Link, Google Scholar|
|6.||Rosenberg JE, Hoffman-Censits J, Powles T, et al: Atezolizumab in patients with locally advanced and metastatic urothelial carcinoma who have progressed following treatment with platinum-based chemotherapy: A single-arm, multicentre, phase 2 trial. Lancet 387:1909-1920, 2016 Crossref, Medline, Google Scholar|
|7.||Balar AV, Galsky MD, Rosenberg JE, et al: Atezolizumab as first-line treatment in cisplatin-ineligible patients with locally advanced and metastatic urothelial carcinoma: A single-arm, multicentre, phase 2 trial. Lancet 389:67-76, 2017 Crossref, Medline, Google Scholar|
|8.||Patel MR, Ellerton J, Infante JR, et al: Avelumab in metastatic urothelial carcinoma after platinum failure (JAVELIN Solid Tumor): Pooled results from two expansion cohorts of an open-label, phase 1 trial. Lancet Oncol 19:51-64, 2018 Crossref, Medline, Google Scholar|
|9.||Powles T, O’Donnell PH, Massard C, et al: Efficacy and safety of durvalumab in locally advanced or metastatic urothelial carcinoma: Updated results from a phase 1/2 open-label study. JAMA Oncol 3:e172411, 2017 Crossref, Medline, Google Scholar|
|10.||Sharma P, Retz M, Siefker-Radtke A, et al: Nivolumab in metastatic urothelial carcinoma after platinum therapy (CheckMate 275): A multicentre, single-arm, phase 2 trial. Lancet Oncol 18:312-322, 2017 Crossref, Medline, Google Scholar|
|11.||Bellmunt J, de Wit R, Vaughn DJ, et al: Pembrolizumab as second-line therapy for advanced urothelial carcinoma. N Engl J Med 376:1015-1026, 2017 Crossref, Medline, Google Scholar|
|12.||Balar AV, Castellano D, O’Donnell PH, et al: First-line pembrolizumab in cisplatin-ineligible patients with locally advanced and unresectable or metastatic urothelial cancer (KEYNOTE-052): A multicentre, single-arm, phase 2 study. Lancet Oncol 18:1483-1492, 2017 Crossref, Medline, Google Scholar|
|13.||Bellmunt J, Choueiri TK, Fougeray R, et al: Prognostic factors in patients with advanced transitional cell carcinoma of the urothelial tract experiencing treatment failure with platinum-containing regimens. J Clin Oncol 28:1850-1855, 2010 Link, Google Scholar|
|14.||Bajorin DF, Dodd PM, Mazumdar M, et al: Long-term survival in metastatic transitional-cell carcinoma and prognostic factors predicting outcome of therapy. J Clin Oncol 17:3173-3181, 1999 Link, Google Scholar|
|15.||Min Lee CK, Li S, Tran DC, et al: Characterization of dermatitis after PD-1/PD-L1 inhibitor therapy and association with multiple oncologic outcomes: A retrospective case-control study. J Am Acad Dermatol 79:1047-1052, 2018 Google Scholar|
|16.||Hua C, Boussemart L, Mateus C, et al: Association of vitiligo with tumor response in patients with metastatic melanoma treated with pembrolizumab. JAMA Dermatol 152:45-51, 2016 Crossref, Medline, Google Scholar|
|17.||Kostine M, Rouxel L, Barnetche T, et al: Rheumatic disorders associated with immune checkpoint inhibitors in patients with cancer-clinical aspects and relationship with tumour response: A single-centre prospective cohort study. Ann Rheum Dis 77:393-398, 2018 Crossref, Medline, Google Scholar|