Is the Current Framework of Academic Oncology Failing Community Providers?

Maurie Markman, MD

The creation of treatment pathways that recognize the chronicity of certain cancers, the complexity of response and resistance, and the volume of accumulated data could represent the first building block toward sustainable decision support tools for community oncologists, according to Maurie Markman, MD.

“If you’re taking care of [patients with] multiple diseases, and the patients come to you and say, ‘What about this? What about that? What is optimal therapy?’ If you’re going from one patient to the next, maybe [you can afford] a 15-minute visit, maybe a 20-minute visit, [but if] a new lesion occurred or a new abnormality was detected on a liquid biopsy and you have 15 to 20 minutes to make decisions, how do you determine all that?”

In this special with OncLive, Markman, who is president of medicine & science at City of Hope Atlanta, Chicago, and Phoenix, discussed the factors that have contributed to the present lack of decision support for community oncologists, including the deficit of financial incentives and insufficiency of clinical trials to serve as much more than a pathway toward regulatory approval.

Why Can’t What’s Broken Be Fixed?

Markman: Why is it so difficult? The rules, the information that we’re generating here [with] trials, they all come from the academic community and, of course, industry. They’re experts in the area, they can focus on [generating that information]. Maybe 100% of their time is focused on just breast cancer or just lung cancer, and that’s full time. It’s dramatically complex with all the things that I noted in the previous [article]. But we know that approximately 85% to 90% of patients with cancer in the United States are taken care of in the community. They’re taken care of by community doctors and community oncologists, who may try to specialize to some extent, and maybe can in a large practice, but the reality is, changes are happening so quickly. 

It’s one thing to say that the academic community can handle this. They can write the papers, they can get up on the podium, they can give continuing medical education programs, but what about the 90% of patients taken care of outside of that academic community? Where is the organized decision support for the community oncologists who are always and only trying to do their best for their patients in this increasingly complex arena? Imagine pathways, gone.

Pathways can provide some value [for] first- and second-line treatment. What about beyond that? What about [the notion that] cancer is a chronic illness? Oftentimes the pathways [only tell you:] here’s one option, here’s a second option, here’s the third option. How does that help you? You have dozens of insurers you deal with who will tell you what you can do or not do, which may not have anything to do with best evidence but instead what they’ll pay for. In your practice, you have to understand that as well. So, [you’re left dealing with] the complexities of the illness, the multiple diseases that most oncologists, certainly in the community, have to deal with, and the fact that there isn’t a real financial incentive for a company to focus on pathways that are robust in a way that will help the community oncologist.

Who’s going to buy that? Who’s going to pay for it? Who’s going to update it, daily, as new things come along? Those are questions that I would submit have not even begun to be addressed. I’m not even sure they’ve adequately been brought up.

Dealing With the Data Dilemma

I want to emphasize the difficulty of even suggesting today that we are at a point where we can put together even minimally adequate decision support tools. We say as an oncology community that we rely on clinical trials; they help us, absolutely. I’m not debating that. I’m not debating that trials give us information about drugs, about their safety and about their efficacy. But the reality is that clinical trials do not [fully] represent the patients we see in the community. Increasingly, they’ll represent the people in the community, [but] age and comorbidities are simply not adequately represented. Then you get into this issue, which is increasingly palpable, of cancer as a chronic illness. So, we are not only talking about first- and second-line treatment, but we’re also talking about third- and fourth-line therapy; we’re talking about patients who may have responded [to treatment] and those with resistance. The complexity is massive.

Clinical trial data do not help [with this]. In general, clinical trials are designed for a purpose that does not relate to what I’m discussing. If you want to get a drug approved, you design a trial that’ll allow you to get the drug approved [in a] homogeneous population. What’s wrong with that? Nothing. But how do we use it in the community? Who is going to pay for it? Who is going to organize [that effort]? No one. National Cancer Institute [NCI] cooperative groups [may have] a little bit [of a role with this], but not much. Also, you’re going to see a lot more, and you have seen a lot more, of these pragmatic trials done outside of this country. That’s a massive problem moving forward.

We need to come up with a structure, a way of looking at these questions, [so we can] address them. What about cancer as a chronic illness? What about sequential vs combination therapy? Of course, all of this is also going to include questions of cost. We’ve got to have a dedicated, robust commitment by someone. I could argue it could be the National Institutes of Health or the NCI [that can] figure out a way to deal with decision support as we move forward. Maybe artificial intelligence [AI] can be a partner here. Maybe AI can help with this. Maybe real-world data that we get from databases, if we shared, can answer these questions for the doctor who is not an expert in the field [and who wants to know, for a] 75-year-old patient with a comorbidity, ‘What’s optimal therapy?’ Where are they going to get that help? We need to, as a society and certainly as an oncology community, start to focus on these issues.

Perspective From a Practiced Gynecologic Medical Oncologist

In [the last article,] I discussed the issues with how rapidly things have changed over the past number of years in the management of cancers from the point of view of systemic therapies. Now, I want to focus on my own area of clinical and research expertise over the past several decades, which has been in gynecologic cancers. About 30 to 40 years ago, we asked the question of how we manage, from a systemic perspective, the three main gynecologic cancers: ovarian cancer, endometrial cancer, and cervix cancer.

Ovarian cancer [has been] a prototype for drug development over many decades. We have the alkylating agents, and of course, the profound change came with the introduction of cisplatin. Then, over 30-plus years, we had cisplatin and carboplatin—the latter of which is a safer, but not better alternative. It’s easier to give and is much better tolerated and has become the standard of care today. Thirty-plus years later, we [got access to] a second agent, which could have been doxorubicin or cyclophosphamide, and then we [had] paclitaxel. Again, we’re talking about 30 years ago. [At that time, we would’ve given] carboplatin and paclitaxel for 6 cycles. If the patient recurred within 6 to 12 months, you’d say they were platinum resistant and not give them platinum. If it was longer than 12 months, you’d give them a platinum again. It was debated whether we should we use 6 months or 12 months [of therapy], but that was it. There were questions about intraperitoneal therapy, which I’ll come back to, and then we used multiple single agents along the line. That was the therapy in the resistance setting. That’s really what you needed to know.

Endometrial cancer was even easier, and one can say distressing at the time, [because] we had essentially no trials. We would give hormonal therapy [with] gestational agents. There was so little work in this area, so we didn’t bother looking for receptors. If it was a low-grade cancer, we would potentially give a hormonal agent. If it wasn’t low grade, and most of them weren’t in the metastatic setting, we would give a variety of drugs just like we gave in ovarian cancer, but of course there was essentially no research in the area. What we used in ovarian cancer, we used in endometrial cancer. The situation was so poor in terms of research and what we knew. It was a period that many people have forgotten, and I tried to forget; we were giving cisplatin, paclitaxel, and doxorubicin together with very little rationale, because so little research was being done.

If we turn to cervix cancer, essentially, we had and still have from the cytotoxic perspective, one drug, and that is platinum. You’d use platinum at a higher dose or lower dose, in combination or alone. When the landmark studies were conducted that demonstrated that chemoradiation improved survival in cervix cancer, we then moved cisplatin up front, giving it weekly with radiation, which was an incredible advance. The issue then was we had no systemic therapy. We had no other drugs that worked in cervix cancer. So, although we might give a variety of agents, we had no therapy that was shown to be of much benefit. That was the state of affairs, until relatively recently. It was not good for patients, but it was easy to remember. One could be trained in this relatively easily––the cookbook was not that complicated because we had limited agents and limited positive trials that influenced our care.

[In the next article, I’ll focus on] just how dramatically things have changed in these three illnesses and the impact of that on decision support and the requirement for decision support.