Abstract
Purpose: Desmoid-type fibromatosis (DF) are locally aggressive neoplasms, with a need for effective systemic treatment in case of progression to avoid the short- and long-term complications of local treatments.
Experimental Design: We retrospectively analyzed the outcomes of adult patients with DF treated with oral vinorelbine (90 mg once weekly) at Gustave Roussy Cancer Institute (Villejuif, Paris, France). Only patients with documented progressive disease according to RECIST v1.1 for more than 3 months (±2 weeks) before treatment initiation were included.
Results: From 2009 to 2019, 90 out of 438 patients with DF were eligible for this analysis. Vinorelbine was given alone in 56 patients (62%), or concomitantly with endocrine therapy in 34 patients, for a median duration of 6.7 months. A partial response was observed in 29% and stable disease in another 57%. With a median follow-up of 52.4 months, the median time to treatment failure (TTF) was not reached. Progression-free rates at 6 and 12 months were 88.7% and 77.5%, respectively. Concomitant endocrine therapy was associated with longer TTF in women [HR, 2.16; 95% confidence interval (CI), 1.06–4.37; P = 0.03). Among 64 patients with documented CTNNB1 mutational status, p.S45F or p.S45P mutations were associated with longer TTF compared with p.T41A or wild-type tumors (HR, 2.78; 95% CI, 1.23–6.27; P = 0.04). Toxicity profile was favorable, without grade 3–4 toxicity, except for one grade 3 neutropenia.
Conclusions: Oral vinorelbine is an effective, affordable, and well-tolerated regimen in patients with advanced, progressive DF. Prolonged activity was observed in patients with tumors harboring CTNNB1 p.S45F or p.S45P mutations.
Translational Relevance
Five percent to 10% of desmoid-type fibromatosis (DF) arise in the context of familial adenomatous polyposis (FAP). The beta-catenin gene (CTNNB1) exon 3 mutations are found in 90% of DF (except those associated with FAP) in two codons, codon 41 (p.T41A) and codon 45 (p.S45F and p.S45P), which are serine and threonine phosphorylation sites required for β-catenin degradation. p.S45F mutations are associated with a higher risk of recurrence. Among 64 patients with DF and documented CTNNB1 mutational status treated with oral weekly vinorelbine for 6 months, p.S45F or p.S45P mutations were associated with longer time to treatment failure compared with p.T41A or wild-type tumors (HR, 2.78; 95% confidence interval, 1.23–6.27; P = 0.04). The functional impact of codons 41 and 45 mutations is certainly different, because the phosphorylation of amino acid 45 targeted by casein kinase-1 serves as a priming phosphorylation for GSK-3β targeting the amino acid 41.
Introduction
Desmoid-type fibromatosis (DF) are monoclonal, fibroblastic proliferative disorders arising in the soft tissues, with a tendency toward local progression and recurrence, but no ability to metastasize (1). The incidence of DF in Western countries is estimated to be five cases per million annually. Five percent to 10% of DFs arise in the context of familial adenomatous polyposis (FAP), and 80% arise in young women (1). The beta-catenin gene (CTNNB1) exon 3 mutations are found in 90% of DFs (except those associated with FAP) in two codons, codon 41 (p.T41A) and codon 45 (p.S45F and p.S45P), which are serine and threonine phosphorylation sites required for β-catenin degradation (2). p.S45F mutations are associated with a higher risk of recurrence after surgery (3).
First-line surgery is no longer the current standard-of-care initial management for DF. Indeed, without any specific treatment, the progression-free rate (PFR) at 5 years is 50% (4, 5). Given the unpredictable natural history of the disease [with the possibility of long-lasting stable disease (SD) and even occasional spontaneous regressions, along with a lack of metastatic potential] and functional problems implied by some anatomic locations, an initial watchful waiting policy can be proposed. This first-line conservative approach is now commonly advised (1). In case of progressive disease (PD), several therapeutic options can be considered and encompass surgery, radiotherapy, cryotherapy, isolated limb perfusion, and systemic treatments (1). A case-by-case discussion in dedicated multidisciplinary tumor boards is encouraged, to take into account the potential loss-of-function and long-term consequences of these treatments (1).
Systemic treatment options include endocrine therapies, NSAIDs, chemotherapy (1, 6), and more recently, multikinase inhibitors (7, 8). The chemotherapy regimens offered to patients with DF encompass anthracyclines, dacarbazine, and other mutagenic agents, or weekly combinations of low dose, intravenous agents such as vinblastine plus methotrexate (1, 6, 9, 10).
Regarding predictive factors for response to systemic treatments, there is no evidence for a role of estrogen or progesterone receptors' expression determined by IHC to predict the efficacy of endocrine therapies. To date, no predictive biomarker has been identified for the response to other systemic treatments. Hence, there is a need for efficient, quality-of-life preserving, and affordable systemic treatments, with limited long-term consequences for the patients.
Vinorelbine is a nonmutagenic cytotoxic agent. Its oral administration at low doses is feasible, active against a broad range of malignancies, and associated with a favorable toxicity profile (11–13).
In this context, we aimed to assess the efficacy and safety of oral vinorelbine in patients with advanced, progressive DF, and to explore the predictive role of CTNNB1 mutations.
Materials and Methods
In 2007, on the basis of the results observed with vinca-alkaloids in DF (14, 15) and the characteristics of the product (Navelbine, Pierre Fabre Oncology), an institutional decision was made to use oral vinorelbine to treat patients with advanced, progressive DF, each time an inclusion in a clinical trial with an investigational drug was not feasible.
We retrospectively reviewed the electronic medical records of these patients, after having obtained the approval of the institutional review board. Because of the retrospective nature of the study, no written consent was needed. The study was conducted in agreement to applicable laws and regulation, and the Declaration of Helsinki.
Patients included in this retrospective study were adults with a diagnosis of DF confirmed by an expert pathologist from the French Sarcoma Group Network RRePS. From 2011, the analysis of CTNNB1 exon 3 mutations was systematically performed initially with Sanger sequencing, and since 2016 with a next-generation sequencing panel with 11 genes from non–small cell lung cancer panel (VELA Diagnostic), according to the manufacturer's recommendations (protocol on demand; ref. 16). Imaging data (CT scan and MRI) were reviewed, and only patients with a documented PD according to RECIST version 1.1 (ref. 17; based on two imaging assessments obtained within less than a 3-month ± 2 weeks interval) preceding the introduction of vinorelbine were included in this analysis.
Demographic data, tumor characteristics (size, location, and CTNNB1 and APC mutational profile), previous treatments, and outcomes (efficacy and toxicity) under vinorelbine were recorded and analyzed. Histopathologic review was performed on hematoxylin and eosin–stained slides from formalin-fixed, paraffine-embedded tumor specimens (biopsy or surgical resection) when available. Two pathologists independently assessed morphologic features considered as related to tumor proliferation (cellular density and presence of interstitial edema; mitotic index was too low to be evaluated in small specimens), immune activation [density of tumor-infiltrating lymphocyte (TIL)], and vascular changes (congestion):
(i) cellular density: presence (1) or absence (0) of areas with high cellularity;
(ii) interstitial edema: presence (1) or absence (0) of focal interstitial edema;
(iii) TILs: mild (0), moderate perivascular (1), or more diffuse (2) infiltration by lymphocytes. The slides were reviewed by two pathologists blinded from clinical data.
Treatment
Oral vinorelbine is usually given at doses ranging from 60 to 80 mg/m2 once weekly (18). However, the apparent clearance of oral vinorelbine is not related to body surface area (18, 19), and the administration of fixed doses is feasible and active against various malignancies (12, 13), with limited pharmacokinetics variability for doses below 120 mg (20). Because most hematologic toxicities are seen for doses of 120 mg and above (20), a flat dose of 90 mg (three capsules of 30 mg each) weekly was chosen. Patients received metoclopramide 10 mg as antiemetic premedication (30 minutes before vinorelbine intake), and received the instruction of taking vinorelbine at the end of a light meal (20). In case of persisting nausea or vomiting, ondansetron 8 mg twice a day was prescribed for 3 days.
If patients had initiated endocrine therapy (tamoxifen and/or LH-RH agonists) before vinorelbine, these treatments could continue at the discretion of the treating physician.
The treatment duration was scheduled to be 6 months. However, when tumor-related pain was persistent after 6 months of vinorelbine, treatment prolongation until symptomatic improvement (as defined below) was considered.
Evaluation criteria
The main evaluation criterion was the response rate (RR) according to RECIST v1.1 (17). Tumor evaluation (using CT scan and/or MRI) was performed every 3 months, or before if clinically indicated.
Secondary evaluation criteria were: PFR at 3, 6, and 12 months; time to treatment failure (TTF); symptomatic improvement; toxicity; and TTF according to tumor CTNNB1 mutational status. TTF was measured from the date of first treatment administration to the date of disease progression, treatment discontinuation due to toxicity, or clinical progression (uncontrolled pain despite optimal use of opioids). Symptomatic improvement was defined as a decrease of at least 50% of the doses of opioids required for tumor-related pain. Toxicities were recorded at each visit, and retrospectively graded according to the NCI-CTC v5.0 (https://ctep.cancer.gov/). Blood cells count, serum chemistry, and liver function tests were monitored monthly for the treatment duration.
Statistical analysis
Descriptive statistics were used to analyze patient characteristics: median, ranges, and 95% confidence intervals (CI). The influence of baseline parameters on RR was evaluated using the χ2 test, Fisher exact test, or Mann–Whitney U test, when appropriate. These parameters were: age, gender, tumor location and size, local and systemic treatments, history of FAP, mutational status (CTNNB1 and APC), and histopathologic features (cellular density, TILs, and interstitial edema). Their influence on TTF was examined in univariate analysis. Survival analyses were performed using the Kaplan–Meier method with the log-rank test. A multivariate analysis was performed using Cox regression. A P < 0.05 was considered statistically significant. Events up to December 31, 2019 were included in the analysis. Calculations were performed with NCSS2020 software (https://www.ncss.com/).
Results
From 2009 to 2019, 438 patients with DF were treated in our expert tertiary cancer center, and 132 received oral vinorelbine. Among these, 90 patients had documented progression before treatment initiation, and were eligible for this analysis.
Median age was 35 years (range, 18–67), 14% of patients had a family history of adenomatous polyposis. The median tumor size was 105 mm (range, 30–290). The median number of prior systemic lines of treatment was 2 (range, 0–5). Patients and disease characteristics are summarized in Table 1. The median duration of follow-up was 52.4 months (range, 2.9–114). Regarding mutational analysis, 63% (42 cases) of 67 samples analyzed exhibited a CTNNB1 exon 3 mutation. Three (4%) were not assessable due to technical failure. Among 22 tumors wild-type for CTTNB1, 13 exhibited an APC mutation. The remaining nine tumors were classified as wild-type for both CTNNB1 and APC, possibly corresponding to a subset of tumors biologically distinct from Wnt-dependent DF (21).
Patients and disease characteristics at the time of treatment initiation (n = 90).
Histopathologic review was performed on 49 available samples. Fifteen samples (31%) exhibited high cellularity, nine (18%) had detectable TILs, and five (10%) had interstitial edema (Table 1).
Treatment disposition
Vinorelbine was given alone in 56 patients (62%), or concomitantly with endocrine therapy in 34 patients (LH-RH agonists in all and associated with tamoxifen in 30 patients).
The median duration of treatment with oral vinorelbine was 6.7 months (range, 2.0–26.2). All patients received at least 3 months (±1 week) of treatment, except two (3%), in whom recurrent grade 2 emesis (despite the use of ondansetron) led to its discontinuation. Both patients subsequently received intravenous vinorelbine (25 mg/m2/week), without recurrence of emesis. In the remaining 88 patients, no dose reduction or interruption was needed.
Treatment outcomes
A partial response (PR) was achieved in 26 patients (29%), an SD in 51 patients (57%), and a PD in 13 patients (14%), as shown in Fig. 1. The PFRs at 6 and 12 months were 88.7% and 77.5%, respectively. The median TTF in the entire population was not reached (Supplementary Fig. S1).
Waterfall plot describing the best response (according to RECIST v1.1) to oral vinorelbine (VNB) in patients receiving concomitant endocrine therapy (ET, n = 34) or not (n = 56). Dotted lines indicate the top and bottom boundaries for SD.
TTF in patients according to CTNNB1 mutational status; wild-type (WT) or codon 41 (p.T41A) versus codon 45 (p.S45F or p.S45P), n = 64.
A shorter TTF was observed in women, median 54.2 months (95% CI, 20.1–68.4) versus not reached in men (HR, 0.40; 95% CI, 0.19–0.82; P = 0.04; Supplementary Fig. S2). The concomitant administration of endocrine therapy had no effect on TTF in the entire population (HR, 1.36; 95% CI, 0.70–2.65; P = 0.38). However, it was associated with longer TTF in women, median not reached versus 31.1 months (95% CI, 14.1–54.2), HR, 2.16 (95% CI, 1.06–4.37; P = 0.03; Cox model interaction test, P = 0.08; Supplementary Fig. S3). No other baseline clinical parameter significantly influenced treatment efficacy. In particular, TTF neither differed in patients with or without FAP (HR, 1.33; 95% CI, 0.52–3.38; P = 0.59), nor according to tumor location (Supplementary Table S1).
Among 64 patients with documented CTNNB1 mutational status, p.S45F or p.S45P mutations (n = 16) were associated with longer TTF (median, not reached) compared with wild-type (n = 22) or p.T41A mutation (n = 26, median TTF, 24.0 months; 95% CI, 16.0–28.0), HR, 2.78; 95% CI, 1.23–6.27; P = 0.04 (Fig. 2).
By multivariate analysis, no independent predictor for TTF was identified. RR and TTF in patients with tumors wild-type for both CTNNB1 and APC did not significantly differ from those observed in other non-p.S45 cases.
Histopathologic features had no significant impact on TTF, but a higher RR was observed in patients with tumors with high-cellular density (60% vs. 18%; HR, 6.66; 95% CI, 1.49–33.81; P < 0.01).
Symptomatic improvement after 3 months of treatment was observed in 53 (74%) of the 70 patients evaluable for pain. Treatment-related toxicities (grade ≥2) are summarized in Table 2. No grade ≥3 toxicity occurred, except for one case of grade 3 neutropenia. Grade ≥2 nausea, asthenia, and neutropenia occurred in 35%, 18%, and 12% of patients, respectively.
Grade ≥2 treatment-related toxicities (according to NCI-CTC v5.0).
Posttreatment outcomes
The 13 patients exhibiting a PD under oral vinorelbine were treated by radiotherapy (n = 5), cryotherapy (n = 3), surgery (n = 2), or other systemic treatments (pazopanib, sorafenib, and liposomal doxorubicin, n = 1 each). After initial disease control under oral vinorelbine, another 14 patients experienced a subsequent PD, after a median time of 11 months (2–20) after vinorelbine discontinuation. In 6 patients, oral vinorelbine was rechallenged for another 6 months, resulting in two PRs and four SDs. The remaining 8 patients received radiotherapy (n = 4), liposomal doxorubicin (n = 2), and cryotherapy (n = 2). All had a PR following these treatments.
Two patients with locally advanced abdominal wall tumors who experienced PR under vinorelbine were subsequently operated (R0) and remain free of disease at the time of the analysis. The remaining 61 patients received no additional treatment. Among them, 11 patients (12%) with SD at the end of the 6 months of treatment with oral vinorelbine subsequently experienced confirmed PR during the follow-up (after a median delay of 12 months, range 4–22), of whom nine (82%) had tumor density or contrast intake decrease on the imaging evaluation at the end of treatment, suggesting that evaluation criteria other than RECIST could be useful in this setting. At the last follow-up, all patients were alive.
Discussion
This study retrospectively assessed the efficacy and safety of oral vinorelbine (90 mg weekly) in a population of adult patients with particularly aggressive forms of DF (with documented progression over 3 months before treatment initiation). We found an objective RR of 29%, and only 14% of patients experienced a PD. The toxicity profile was favorable, with only 2 patients (2%) having discontinued the treatment due to toxicity. Besides, the treatment was active in FAP-associated DF.
Among conventional chemotherapy agents active against DF, anthracyclines are active (6, 22), but may cause life-threatening toxicities (cardiac dysfunction and secondary malignancies), making them poorly attractive in patients with a nonmetastasizing disease such as DF. Our institutional policy is to consider their use only in patients with potentially life-threatening locations (mesenteric or cervical DF), alone or in combination with locoregional treatments.
Another regimen with more acceptable toxicity profile is the combination of vinblastine and methotrexate, which displays clinical activity in DF (9, 10, 14). Importantly, this regimen is administrated weekly by intravenous route, usually for 1 year, making its acceptance by patients a serious concern.
Regarding endocrine therapy, its activity was observed in a population of patients that was not selected on the basis of rapid progression (23). One could, therefore, argue that a number of SDs and PRs observed in this setting might be due to the natural history of the disease rather than the antitumor activity of endocrine therapy per se. However, we observed shorter TTF in women on one hand, and longer TTF in women receiving endocrine therapy concomitantly to oral vinorelbine on the other hand, suggesting a potential role of hormonal factors in long-term control of the disease. Conversely, we observed a plateau in TTF in male patients after 18 months (Supplementary Fig. S2).
Regarding another drug, phase II studies (24–26) have shown a meaningful activity of imatinib in DF. However, imatinib is not approved in this indication, and its cost remains a potential barrier to its wide use in DF.
More recently, two studies have shown the efficacy of multikinase inhibitors [sorafenib 400 mg every day (ref. 7) and pazopanib 800 mg every day (ref. 8)] in DF. However, the populations of patients were particularly heterogeneous. Indeed, progressive DF, unresectable tumors, and stable but symptomatic tumors were included in the sorafenib trial, whereas DF progressive over 6 months with 72% of asymptomatic patients (PS 0) were included in the pazopanib trial.
In both studies, up to 20% of patients discontinued treatment because of treatment-related toxicities, and treatment duration was prolonged (1 year or longer; refs. 7, 8). Hence, the optimal duration of treatment with sorafenib or pazopanib remains to be determined. Given the potential toxicity and the cost associated with the use of these agents, we consider that other treatment options are needed. Finally, data on the activity of sorafenib and pazopanib in patients with FAP-associated DF have not been reported to date.
In this context, we consider that oral vinorelbine represents an attractive therapeutic option in view of its activity, its favorable toxicity profile, the oral route of administration making its use more convenient for patients, and its relatively low cost compared with other oral agents. Indeed, in the French setting, oral vinorelbine treatment during 6 months presents a cost of €2,632, favorably comparing with other treatment regimen options like liposomal doxorubicin (six cycles; ref. 1) or sorafenib (24 months; ref. 7) with a cost of €11,425 and €74,521, respectively (Supplementary Table S2). To confirm this finding, we have also conducted a preliminary estimation of sequential strategies cost within a theoretical cohort of 100 patients (Supplementary Table S3) that tends to corroborate this trend; even more robust economic evidence is needed (in France and in other jurisdictions), with specific cost-effectiveness studies. Moreover, to date, oral vinorelbine is available in most European countries, but not in the United States or in Australia.
Finally, few studies have investigated the association of CTNNB1 mutational status with response to systemic treatments. Imatinib seemed to induce more progression arrests in tumors with mutations in codon 45 (p.S45) in a series of 34 patients (27). In contrast, p.S45F mutation was associated with poor responses to the cyclooxygenase-2 inhibitor, meloxicam, in another series of 33 patients (28). Data obtained from the sorafenib and pazopanib trials have not been published to date. In this series, mutations in codon 45 (p.S45) were detected in 18 (27%) of the 67 samples assessed for CTNNB1 mutational status, and were associated as a predictive marker with prolonged disease control. Keeping in mind that p.S45F mutations are associated with a higher risk of recurrence after surgery (3), possibly due to a larger size at diagnosis (29), these results question the predictive role of CTNNB1 mutations, and will require confirmation in independent cohorts. One could hypothesize that rapidly growing tumors could be more sensitive to mitotic spindle poisons such as vinorelbine, consistent with our clinicopathologic findings of higher RRs in tumors with high-cellular density.
The functional impact of these two codons is certainly different, because the phosphorylation of amino acid 45 targeted by casein kinase-1 serves as a priming phosphorylation for GSK-3β targeting the amino acid 41 (30, 31).
Overall, this study is the largest (n = 90) ever reported on systemic treatments in DF. The conclusions are strengthened by the homogeneity of the patients' population treated in a large referral center, and the prolonged follow-up. However, the retrospective nature of the study is a clear limitation. Although only patients with rapidly progressive DF were included, a placebo-controlled study would have been preferable, given the natural history of DF and possible spontaneous regression or stabilization of the tumor. As an illustration, an RR of 20% was reported in a recent placebo-controlled study (7). Likewise, in a retrospective series of 74 patients with FAP-associated DF, spontaneous stabilization or regression was reported in 20 cases (27%; ref. 32).
Hence, we do not know whether the delayed responses we observed after vinorelbine discontinuation were related to the treatment or to spontaneous regression of DF. In addition, we could neither evaluate the impact of genomic features in all patients, nor assess treatment adherence, a major issue with oral anticancer agents (33).
Conclusion
Oral vinorelbine (90 mg flat-dose, once weekly) for 6 months is active in patients with progressive DF, with a disease control rate of 86%. This treatment is affordable and has low toxicity. Prolonged disease control was more frequent in men and in patients with tumors harboring CTNNB1 mutations in codon 45 (p.S45F or p.S45P). These findings might serve as a benchmark for future trials assessing new agents in progressive DF.
Disclosure of Potential Conflicts of Interest
O. Mir reports personal fees from Amgen, AstraZeneca, Bayer, Blueprint Medicines, Bristol-Myers Squibb, Eli Lilly, Ipsen, Lundbeck, Merck Sharpe & Dohme, Pfizer, Roche, Servier, and Vifor Pharma outside the submitted work, as well as reports being a shareholder of Amplitude Surgical, Ipsen, and Transgene. M. Faron reports personal fees from HRA Pharma (board) and Novartis (travel grant) and nonfinancial support from Ipsen (travel grant) and Pfizer (travel grant) outside the submitted work. C. Le Péchoux reports other from AstraZeneca (honoraria for advisory board paid to institution), Amgen (honoraria for educational lecture paid to institution), Nanobiotix (honoraria for advisory board paid to institution), Roche (honoraria for advisory board paid to institution), Medscape (honoraria for presentation paid to institution), and Lilly (honoraria for educational lecture paid to institution) and personal fees from prIME Oncology (honoraria for teaching session) outside the submitted work. E. Rouleau reports grants and personal fees from AstraZeneca and BMS, and grants from Roche outside the submitted work. J. Adam reports personal fees from AstraZeneca, Bayer, Bristol-Myers Squibb, and Roche and grants and personal fees from MSD outside the submitted work. A. Le Cesne reports personal fees from PharmaMar, Bayer, Lilly, and Deciphera outside the submitted work. No potential conflicts of interest were disclosed by the other authors.
Authors' Contributions
O. Mir: Conceptualization, data curation, formal analysis, supervision, validation, investigation, writing-original draft, project administration, writing-review and editing. C. Honoré: Conceptualization, formal analysis, investigation, writing-review and editing. A.N. Chamseddine: Conceptualization, investigation, methodology, writing-review and editing. J. Dômont: Conceptualization, investigation, writing-review and editing. S.N. Dumont: Conceptualization, investigation, writing-review and editing. A. Cavalcanti: Conceptualization, investigation, writing-review and editing. M. Faron: Conceptualization, investigation, methodology, writing-review and editing. F. Rimareix: Conceptualization, investigation, writing-review and editing. L. Haddag-Miliani: Conceptualization, formal analysis, investigation, writing-review and editing. C. Le Péchoux: Investigation, writing-review and editing. A. Levy: Conceptualization, investigation, writing-review and editing. C. Court: Investigation, writing-review and editing. S. Briand: Investigation, writing-review and editing. E. Fadel: Investigation, writing-review and editing. O. Mercier: Investigation, writing-review and editing. A. Bayle: Formal analysis, methodology, writing-review and editing. A. Brunet: Formal analysis, writing-review and editing. C. Ngo: Formal analysis, investigation, writing-review and editing. E. Rouleau: Conceptualization, formal analysis, investigation, writing-review and editing. J. Adam: Conceptualization, formal analysis, investigation, writing-review and editing. A. Le Cesne: Conceptualization, investigation, writing-review and editing.
Acknowledgments
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Footnotes
Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).
Prior presentation: Presented at the ASCO Annual Meeting 2016 (abstract 11050).
Clin Cancer Res 2020;26:6277–83
- Received May 13, 2020.
- Revision received July 15, 2020.
- Accepted August 28, 2020.
- Published first September 1, 2020.
- ©2020 American Association for Cancer Research.
References
Volume 26, Issue 23
