COVID-19 Sharpens Focus on Drug Holidays

Howard A. “Skip” Burris, MD

Having patients go on drug holidays or intermittent therapy is a common strategy during treatment of prostate cancer and several other malignancies, whether to provide temporary relief from adverse effects (AEs) influencing quality of life (QOL), to prevent long-term health impacts, or to lessen drug resistance that would render the therapy ineffective. Now, however, the coronavirus disease 2019 (COVID-19) pandemic has made the decision to take a break from ongoing therapy more complicated and more urgent.

Before the pandemic, “there was a lot of anxiety from different patient groups” about interrupting their therapy, said Howard A. “Skip” Burris, MD, chief medical officer and president of clinical operations at Sarah Cannon Research Institute in Nashville, Tennessee, and a 2014 Giants of Cancer Care® award winner in the Drug Development category.

“I had some men who were older, in their 70s, who were worried about doing intermittent androgen therapy, and yet when COVID hit and they were more worried about coming to the clinic, they put themselves on a break,” Burris said in an interview with OncologyLive®. “I had women who were on maintenance trastuzumab [Herceptin] in remission for a long time who were worried about coming to the clinic, so they took their first break.”

The American Society of Clinical Oncology (ASCO) Survey on COVID-19 in Oncology Registry (ASCO Registry) is collecting data on how the epidemic is affecting the delivery of care and patient outcomes, including for those who paused treatment or switched therapies, noted Burris, who served as ASCO president from 2019 to 2020. “Everybody in the ASCO Registry has to be on treatment, either metastatic or adjuvant, or they have to have been within a year of completing therapies,” he said. “I think we’re going to learn a lot from these data.”

The data could prove illuminating in part because, as common as drug holidays are, their safety and efficacy varies in trials. Data are often contradictory, equivocal, or simply lacking for many scenarios, and it can be unclear why some patients benefit from therapy pauses while others experience progression.

Several research teams are now testing adaptive therapies that include holidays and dose modulation based on disease biomarkers or evolving tumor cell resistance, rather than uniform sequencing plans.

Regarding the definition of a “drug holiday,” Burris said the term means different things to different people. It can mean giving maintenance therapy intermittently, such as stopping chemotherapy and continuing only a checkpoint inhibitor administered once every 3 weeks, or stopping therapy in favor of active surveillance when a patient appears to be in complete or partial remission.

Jacob Scott, MD, DPhil

However, it can be difficult to gauge whether a patient still has active disease, and he said that he’s eagerly following adaptive therapy trials and other research that aims to inform intermittent dosing strategies.

“For the individual patient with cancer and metastatic disease, we’re closer to treating them as if they had a chronic disease. We need to then start thinking about alternative approaches, whether it be drug holidays, sequencing therapies, or whatever the terminology is,” he said. “We’ve had enough success with these drugs, and we’ve got patients living longer but not yet cured.”

Prostate Cancer Questions

In prostate cancer, “tons of studies in the past [have been] done on intermittent androgen deprivation [IAD]. Those roundly failed,” said Jacob Scott, MD, DPhil, a radiation oncologist at Cleveland Clinic Taussig Cancer Center in Ohio who studies evolutionary oncology. “One schedule doesn’t fit every person, and it might even be that the schedule that works for patient X stops working for patient X in 6 months because the cancer has changed. So it’s really about trying to understand the dynamics of an individual person’s cancer from some observables that we have.”

IAD may be the most thoroughly studied use of drug holidays. Continuous androgen deprivation (CAD) is the standard therapy for metastatic hormone-sensitive prostate cancer, but the resulting drop in serum testosterone levels is associated with sexual dysfunction, risk of cardiovascular disease, loss of bone and muscle strength, cognitive effects, and other AEs.¹

Stopping androgen therapy for some period of time to allow testosterone levels
to rise again is an important option for many men, said Nicholas J. Vogelzang,
MD, a clinical professor at University of Nevada School of Medicine in Las Vegas and a 2018 Giants of Cancer Care® winner in Genitourinary Cancers.

Prostate cancer treatment “is a bit of a rollercoaster ride for these patients, and they look forward to being off hormone therapy,” Vogelzang said. “I may be whistling in the dark, but I really hope and pray that for some of these men, they can get long periods of time when their testosterone can recover so they will be able to remain relatively normal. But I am regularly disappointed.”

Findings from the largest IAD trial (N = 3040) were published in 2013 by Maha H. A. Hussain, MBChB, a 2015 Giants of Cancer Care^® award winner, and colleagues, including Vogelzang. Investigators found significantly better erectile function and mental health (P < .001 and P = .003, respectively) with intermittent therapy than with continuous therapy 3 months after randomization but not thereafter. Additionally, the overall incidence of grade 3 or 4 AEs was similar between the 2 therapeutic approaches: 30.4% for IAD and 32.7% for CAD.²

Across the study population, median overall survival (OS) was higher in the continuous group at 5.8 years compared with 5.1 years in the intermittent group, representing a 10% relative increase in the risk of death with intermittent therapy (HR for death with intermittent, adjusted for stratification factors, 1.10; 90% CI, 0.99-1.23). The results “failed to show that intermittent therapy was noninferior to continuous therapy with respect to survival,” the investigators wrote. Additionally, they noted, findings suggested that intermittent therapy could hold
a 20% greater risk of death, but there were too few events to make a determination.²

Subsequent analyses have reached varying conclusions. A 2015 study
of data from 15 trials found support for the noninferiority of IAD, no difference
in OS for the 2 groups, and minimal differences in patients’ self-reported QOL, although there was some improvement in physical and sexual functioning with IAD.³

Another 2015 study, by Hussain, Vogelzang, and colleagues, examined findings from 7 phase 3 trials and concluded that none demonstrated survival superiority of IAD compared with CAD. The trials that found IAD was not inferior were based on wide, prespecified noninferiority margins that likely would not be considered comparable by doctors or patients, they wrote.⁴

Last year Hussain et al said findings from another meta-analysis showed the paradigm for IAD has “not lived up to expectations.”⁵ However, they noted that newer treatment regimens, such as CAD combined with docetaxel and abiraterone acetate (Zytiga) in patients with metastatic hormone-sensitive prostate cancer, have allowed patients with prostate cancer to live longer, making the impact of AEs and the need for deintensification strategies more prominent.

Other recent options include FDA approvals of the second-generation androgen receptor antagonists apalutamide (Erleada) and enzalutamide (Xtandi) in combination with androgen deprivation therapy (ADT) in metastatic castration-sensitive settings and darolutamide (Nubeqa) plus ADT in nonmetastatic castration-resistant disease.⁶

Hussain and colleagues acknowledged that experienced clinicians recognize the benefits of IAD but that doctors and patients must understand its potential negative impact, especially given the lack of data on intermittent use of newer therapies.⁵

The trials examined in the recent meta-analysis used several thresholds for pausing ADT: after 6 months; when prostate-specific antigen (PSA) levels fall below 4 ng/mL or reach 1 ng/ mL or lower; or when PSA decreases 80% or 90% below baseline levels.

Vogelzang said he looks for patients to possibly take a break after 18 months if their PSA is below 0.1 ng/mL. “When do you stop? The answer is, no one knows,” Vogelzang said. “I have arbitrarily taken the point that maybe 18 months is enough. That’s just my personal world view.”

Exploring Adaptive Dosing

Investigators at Moffitt Cancer Center in Tampa, Florida, have tried to lengthen time
to progression (TTP) by personalizing on-and off-treatment cycles in a way that slows the evolution of cellular resistance to ADT. Investigators developed a game theory model based on interactions of intratumoral subpopulations with testosterone.⁷ When therapy is administered continuously at the maximum tolerated dose (MTD), investigators theorize, there is a rapid expansion of resistant T cells, whereas adaptive dosing allows the responsive cell populations to recover (Figure 7).^7,8

In a small trial, men with asymptomatic or minimally symptomatic metastatic castration-resistant prostate cancer (mCRPC) were enrolled after they achieved a PSA reduction of 50% or greater with abiraterone as frontline therapy.^8,9

The study group comprised 15 men who were off the drug for at least 3 months before it was restarted for PSA progression. After more than 11 months of follow-up, 7 of the 15 had completed at least 2 adaptive therapy cycles at data cutoff in January 2019.⁹

Alexander Anderson, PhD

Investigators calculated that the median radiographic progression-free survival (rPFS) would be no less than 30 months (P = .0068, Fisher exact test). Investigators said that result compared favorably with the 16.5-month median rPFS in the COU-AA-302 trial (NCT00887198) of abiraterone plus prednisone versus placebo in mCRPC.^8,9 The Moffitt trial also required much less of the drug, with patients using only 49% as much abiraterone as with continuous therapy.

Such adaptive strategies exploit competition between treatment-resistant and treatment-sensitive tumor cells. Rather than using a traditional MTD strategy that aims to eradicate as many cells as possible at the start, the adaptive therapy allows some sensitive cells to remain and “out compete” resistant cells, slowing their growth during treatment breaks. Although resistant cells eventually increase enough that the treatment fails, investigators say that in a number of cancers, this type of tumor control strategy extends TTP longer than MTD does.

“The crucial point here is, they’re doing dose modulation or holidays in response to how the tumor grows,” said Alexander Anderson, PhD, chair of the Integrated Mathematical Oncology department and senior member at Moffitt Cancer Center. “If you don’t see much growth or it doesn’t grow too quickly or return to the size it was within a period of time, they’re not treating or they’re continuing to deescalate the dose. When it eventually returns to the size it started at, that’s when they kick the treatment back in.”

Uniform Dosing Hits Snags

That personalized approach contrasts with previous drug-holiday trials that randomly assigned patients to uniform dosing schedules, not only in prostate but also breast, melanoma, colorectal, and other cancers.

Findings from one such trial were presented at the European Society for Medical Oncology Breast Cancer Congress 2019 in Berlin, Germany.¹⁰ The phase 3 STOP&GO trial (EudraCT 2010-021519-18; BOOG 2010-02) enrolled 420 patients with advanced HER2-negative breast cancer and randomized them to either 8 cycles of continuous chemotherapy or an intermittent schedule of 4 cycles, a break, and then 4 more cycles.

The combined median OS for patients receiving first-line therapy of paclitaxel plus bevacizumab (Avastin) or second-line treatment with capecitabine or nonpegylated liposomal doxorubicin was 20.3 months for those on intermittent treatment compared with 23 months for those on continuous treatment (HR, 1.93; 95% CI, 1.26-2.95).¹⁰ A separate analysis found that physical QOL declined more in the intermittent arm and that mental QOL improved slightly in both groups, but increased more in the continuous arm.¹¹

Serial holidays in melanoma treatment were explored in a phase 2 trial (NCT02196181) presented at the American Association for Cancer Research Virtual Annual Meeting 2020. The trial enrolled 242 patients with advanced BRAF V600E/K– mutated disease. After receiving dabrafenib (Tafinlar) plus trametinib (Mekinist) for 8 weeks, 206 nonprogressing patients were randomized to continuous dosing with the combination or intermittent treatment on a 3-week-off, 5-week-on schedule. Median progression-free survival (PFS) was 9 months with continuous dosing versus 5.5 months with intermittent (P = .064). Median OS was 29.2 months in both groups (P = .93) at a median follow-up of 2 years.¹²

Intermittent therapies have also been studied as ways to improve QOL for patients with metastatic colorectal cancer (mCRC). The OPTIMOX1 trial (NCT01023633), for example, randomly assigned patients to 2 different regimens of leucovorin and fluorouracil with oxaliplatin (FOLFOX). Arm A received continuous standard FOLFOX4 every 2 weeks until progression; arm B received FOLFOX7, a simplified leucovorin and fluorouracil regimen with high-dose oxaliplatin, for 6 cycles, followed by 12 cycles without oxaliplatin, and then FOLFOX7 again.¹³

The 2 groups fared similarly: Median PFS was 9.0 months in arm A and 8.7 months in arm B, and median OS was 19.3 and 21.2 months, respectively. Arm B had somewhat fewer grade 3/4 toxicities (54.4% arm A vs 48.7% arm B). The investigators concluded that oxaliplatin can be safely stopped after 6 cycles in a FOLFOX regimen.¹³

A 2013 review of mCRC rechallenge strategies discussed OPTIMOX1 and several other trials that had treatment holidays of a sort, but the overall implications were unclear due to varying study designs and results. The studies included the phase 3 COIN trial (ISRCTN27286448), which tested preplanned treatment holidays in 1630 patients with advanced colorectal cancer.¹⁴ Patients were randomized to receive either a continuous oxaliplatin-fluoropyrimidine combination (arm A), continuous chemotherapy plus cetuximab (Erbitux; arm B), or intermittent chemotherapy (arm C).¹⁴

In comparing results from arms A and C, investigators found that patients in the per-protocol population who received continuous therapy had a median OS of 19.6 months (interquartile range [IQR], 13.0-28.1) versus 18.0 months (IQR, 12.1-29.3) for those who had intermittent treatment. In terms of AEs, those in the continuous therapy arm experienced lower rates of nausea and vomiting (2% vs 8%) than those on intermittent therapy, respectively, but higher rates of grade 3 or worse hand-foot syndrome (4% vs 3%) and peripheral neuropathy (27% vs 5%).¹⁵

In their review of multiple studies, Tonini et al concluded that the data do not clearly define an optimal role for a treatment holiday in which all therapy is paused and for chemotherapy-free intervals with maintenance. They did see the potential for a role for rechallenge therapy for fit patients in third-line or fourth-line mCRC treatment. “Intermittent treatment could be an important strategy in management of [patients with] mCRC when there is not the purpose of gaining an important tumor shrinkage,” they wrote.¹⁴

Intermittent immunotherapy has also been studied. One small phase 2 trial (NCT03126331) used a partially personalized strategy. Fourteen patients with metastatic renal cell carcinoma who received prior antiangiogenic therapy were treated with nivolumab (Opdivo) for 12 weeks. Of those, 5 achieved a 10% or greater reduction in tumor burden and entered a treatment-free observation phase with reimaging every 12 weeks. After a median 48 weeks’ follow-up, 1 patient had an increase in tumor burden of 10% or greater and restarted therapy. The remaining 4 patients had a sustained response for a median of 34 weeks (range, 16-53) off therapy.¹⁶

Future Strategies

Figure. Evolution of Cancer Cell Population in Standard Versus Adaptive Dosing

Moving forward, investigators should seek to tailor therapy according to tumor composition rather than using one-size-fits-all schedules, holidays, and doses, Anderson and colleagues at Moffitt showed in a preclinical, mathematical modeling study.¹⁷

They found that tumors made up of a mix of sensitive and resistant cells, such as breast cancers and melanomas, were best managed by an adaptive approach that exploits the fitness cost of resistance and cellular spatial dynamics. However, homogeneous tumors composed of sensitive cells, such as testicular cancer and some lymphomas, tend to respond better to MTD. Translating these findings to the clinic would require measuring initial variation in drug sensitivity using immunohistochemistry and then periodically assessing of tumor burden using systemic biomarkers, circulating tumor cells, cell-free DNA, or imaging, the investigators said.¹⁷

Other preclinical studies by the Moffitt group showed how adaptive therapies could work in breast cancer. Investigators applied paclitaxel therapy to cell lines of metastatic triple-negative and estrogen receptor (ER)–positive breast cancer and measured tumor response with MRI. After initial intensive therapy achieved tumor control, it was maintained with progressively smaller doses. In 60% to 80% of animals, no treatment was necessary for intervals as long as several weeks.¹⁸

A subsequent study investigated anti-estrogen therapy in which ER-positive xenografts were treated with continuous tamoxifen and cycles of estrogen suppression and stimulation. Brief interruptions in drug administration provided equal tumor control while using up to 50% less drug and maintaining high expression levels of ER and lower levels of MDR1, a glycoprotein.

One focus of adaptive therapy research has been tyrosine kinase inhibitors (TKIs), which inevitably stop working as resistance is acquired. Studies have demonstrated that some patients with non–small cell lung cancer (NSCLC) benefit from re-treatment with the same EGFR-targeted TKI after a holiday.²⁰ However, preclinical work by Scott and colleagues revealed patterns or “maps” of collateral sensitivities that displayed cross-resistance to first-line ALK TKIs in ALK-positive NSCLC, including after drug holidays, suggesting other strategies might provide optimal results in second-line settings.²¹ Another study tested resistance to a battery of drugs in Ewing sarcoma cells after chemotherapy and again found replicable patterns of sensitivity and resistance.²²

Due to the randomness of cell evolution, having sensitivity maps does not on its own allow predictions in individual patients, Scott said, adding that with his next study, of EGFR-positive lung cancer cells, he hoped to overcome that gap by gathering tumor genomic data linked to sensitivity patterns. “Once you have that map, the underlying hypothesis is that it’s fundamental.

Therefore, when a patient comes to you, you could biopsy their tumor, do next-generation sequencing, and then match that sequencing result up to your map of in vitro genotype-phenotype mapping.”

Those genetically informed maps could someday allow continuous therapy adjustments over the course of treatment, he said. That could prove useful not only during regularly scheduled care but also after unexpected breaks, such as those that have occurred during the COVID-19 pandemic. Patients would “come in and tell you, ‘Whoops, I took a month off,’ and you say, ‘OK, well, we need to reassess the state of the tumor,’ with the assumption that it has kind of wandered around evolutionarily a little bit. As soon as you know where it starts, you can calculate the future to a certain extent,” Scott said.

Burris noted that frequent next-generation sequencing testing is already technologically feasible but not practically accessible to many patients. “I’m hopeful that we will have price points around DNA sequencing,” he said. “That you could do sequential liquid biopsies, sequential blood-based sequencing, and try to see what patterns of resistance are emerging.”