Joan was discharged from the hospital 24 days later after her blood counts started to recover, which indicated that Lars’s bone marrow had engrafted within her. She had bouts of graft versus host disease over the ensuing months, and a couple of fevers that landed her back in the hospital. One of those fevers—caused by respiratory syncytial virus, or RSV, the virus that leads to croup in children, but which can be deadly to a person with no immune system—landed her in the ICU again. But she remained in remission three years and counting following her transplant. Playing the odds, she was more likely cured than not.
She returned to work about a year after her diagnosis. The first time she scrubbed back into the OR for a case (another hernia repair!), a round of applause greeted her, led by the surgeon she had worked with for so long. He had visited her a few times during her multiple stays at the hotel Cleveland Clinic. She may even have seen his eyes tear up above his blue surgical mask.
Cure is a funny word within cancer circles. For people with leukemia, it is defined as having gone five years without the leukemia returning. But most oncologists I know, and patients too, avoid using the word with an almost religious or superstitious fervor.1 A few times, earlier in my career, I told patients they were cured after we celebrated their five-year follow-up visit. Months later, a couple of them returned to my clinic with new laboratory abnormalities, and bone marrows that showed leukemia precursor diseases, such as myelodysplastic syndrome. Perhaps the chemotherapy we used to treat their leukemias knocked their bone marrows back to a primordial, pre-leukemia state that was now rearing its ugly head again. Perhaps the chemotherapy itself had caused the bone marrow changes. Regardless of the scenario, I became much more cautious in using the word.
I think we all fear that if we declare a person cured, we will be tempting the gods, and might accidently raise the malignant golem from its torpor. So instead, our focus is on phrases like “continued remission,” or “durable remission.” And we knock on a lot of particleboard tabletops for good luck when we use them.
The treatment of acute leukemia has become much more sophisticated, ranging from the use of new drugs that are “personalized” to attack the distinct genetic basis of the leukemia an individual has developed, to the incorporation of our new understanding of the genetic basis of leukemia so we can predict who would most likely benefit from a bone marrow transplant before the leukemia relapses.
The US Food and Drug Administration has just approved drugs like midostaurin and gilteritinib, which target a genetic abnormality called FLT3, which is present in about one-third of AML patients, because they extend people’s lives.2 Similarly, the drug David received briefly, enasidinib, and ivosidenib target the genetic mutations IDH2 and IDH1 (found in about 15 percent of people with AML) and in so doing induce responses in about one-third of patients whose AML has relapsed.3 These drugs are also being incorporated into treatment after a bone marrow transplant, to prevent the leukemia from returning.
But what is available for the other 52 percent of AML patients who don’t have a FLT3 or IDH abnormality?
Other drugs in development are targeting additional genetic mutations. These mutations include TP53, which drives a leukemia that is exquisitely resilient to chemotherapy and portends particularly poor outcomes, as well as SF3B1, or a variety of other genetic mutations that affect a small percentage of leukemia patients and are similarly named with mixes of letters and numbers.4
My dad, a former newspaper reporter and no slouch in the “knows how to doggedly ask questions” department, once grilled me with a line of inquiries that boiled down to “Why can’t researchers find a cure for cancer?”
Never mind cancer as a whole—leukemia itself is not one cancer, but dozens of cancers, each with its own distinct genetic fingerprint. Given the genetic complexity of leukemia, the future of therapeutic approaches will involve the need for many drugs, each of which targets a different genetic component of leukemia specific to just a small percentage of patients, and likely used in conjunction with standard chemotherapy.
Alternatively, there is immunotherapy, which takes advantage of a patient’s own immune system to combat cancer. Being broadly similar in principle to the primary effect of bone marrow transplantation, it is being studied extensively in acute lymphocytic leukemia.5 One approach involves chimeric antigen receptor (CAR) T-Cells, in which the immune system is deliberately exposed to a patient’s particular leukemia outside the body to target that leukemia. The immune system portion targeting the leukemia is allowed to expand, and is reinfused into the body to attack the leukemia where it lives. Specific CAR T-Cells have been approved for lymphomas and for children, adolescents, and young adults with acute lymphocytic leukemia. Their response rates in these patient populations, many of whom had no other treatment options left, have been remarkable, even nearing 90 percent, with 50 percent of patients remaining in remission one year after they receive their immune cell infusion. Whether they work as well in myeloid cancers, such as acute myeloid leukemia, is being explored.
Sarah’s BCR-ABL levels continued to decrease, crossing the magical Rubicon of being less than 0.1 percent. They never disappeared entirely, though, so stopping the imatinib was not in the cards. Years later, she started her own business and found time on top of that to join Joey’s school’s PTO. She rarely, if ever, missed a dose of the chemotherapy, and she managed to keep her drinking in check. Joey grew to be a healthy boy, his exposure to the imatinib in utero a non-issue. Although Sarah, still spirited as ever, blamed the drug for his being a “royal pain-in-the-ass.”
Because we’ve seen so much success treating CML, given the effectiveness of imatinib and other tyrosine kinase inhibitor drugs, in some ways we leukemia caregivers and researchers have the luxury of focusing on more nuanced therapy questions: Can we stop treatment in patients whose PCR for the BCR-ABL mutation of the Philadelphia chromosome has gone negative? How negative is negative enough: A test that can detect 1 in 1,000 cells with the BCR-ABL, or 1 in 50,000? Can we reduce the dose of imatinib’s children and grandchildren without losing efficacy? How can we help the minority of patients who don’t fare well on the usual tyrosine kinase inhibitors?
When imatinib and its children or grandchildren don’t work, it’s usually because a person’s CML has acquired an additional mutation that prevents the tyrosine kinase inhibitor drug from fitting neatly in the “pocket” of the tyrosine kinase protein that it is supposed to inhibit. A common mutation—and yes, its name comprises another series of letters and numbers—is T315I. Some grandchildren and even great-grandchildren of imatinib have been developed to help patients with just such a resistant CML. But with the improved efficacy comes additional side effects, such as a higher risk of forming blood clots, a risk increasingly detected as more and more CML patients, whose lives used to be measured in a handful of years, become decades-long survivors. Drug development in CML is focusing more on those rare patients for whom the approved tyrosine kinase inhibitors don’t work, and in recognizing side effects in long-term survivors.6
David was buried in Ashtabula, in the same cemetery as his parents and grandparents were interred. More than a hundred people attended the service. Eric gave a moving eulogy on how hard it was to finally let his father go, and how his dad only slipped away after he and Susan gave him permission to do so.
I’ve heard palliative care and hospice nurses refer to the transition stage of death, similar to the transition stage of birth, usually the point at which a woman declares, “GET THIS BABY OUT!!!!” in no uncertain terms. The path to dying can be equally treacherous, equally painful. Having your children reassure you that you don’t have to worry about them, that they will be okay without you, can make that journey more bearable.
At the unveiling of the gravestone, weeks later, Susan stuck a small Indians flag in the ground by his grave, along with a “We’re #1!” Indians foam finger. His was not the only plot in the cemetery with those memorabilia.
Older adults with AML are probably the most challenging group of patients to treat, in large part because the biology and genetics of their cancers is even more complex than that in younger adults. Consequently, “one size fits all” treatment approaches are least likely to work, particularly as older patients tend not to have the “good risk” types of leukemia (as fraught with irony as that term is). Joan’s acute promyelocytic leukemia is an example.
But what if we could identify which of these patients would benefit from more intensive therapy, and which from less intensive, outpatient therapy, at the time of their diagnosis? One approach to doing this lies in the power of computers to analyze pathologic, clinical, and genetic data using what’s called “machine learning” or “artificial intelligence (AI)” programming.
AI has been used in remarkable ways in medicine in the past few years. One program has gotten so sophisticated at distinguishing cancerous skin lesions (such as melanomas) from benign abnormalities that it can outperform dermatologists.7 Similar pattern recognition software is being applied to bone marrow samples to determine if machines are better at diagnosing myelodysplastic syndromes or leukemias than pathologists, or if machines can assist pathologists in making diagnoses. It’s hard for me to believe any machine can outperform Karl, but this is admittedly an engrained, immutable bias of mine.
AI has already been used in assessing prognosis in myelodysplastic syndrome and has outperformed other prognostic schemas, given its ability to be applied at multiple time points in a patient’s disease course.8 In the same disease, it can predict which patients are likely to get better when treated with a particular chemotherapy, and which have cancers resistant to that chemo. If only we had known that prior to starting David on a weeks-long hospital course that was unsuccessful.
Outcomes in leukemia also should be placed in context of where we are now, two decades into the new millennium, and where we’ve been. In 1975, all comers with leukemia, including those with both chronic and acute varieties, had a 33 percent chance of being alive five years after a diagnosis. That survival rate has doubled, to 66 percent. With the rapidity of scientific discoveries and the incorporation of genetics and computer-based approaches accelerating as they have over the past few years, I anticipate those survival rates will increase dramatically during the next decade.
After all, the goal for me and for my clinical and research colleagues is to put ourselves out of a job as quickly as possible.