Discussion
These data demonstrate a striking increase in reported pediatric cancer cases in Florida over the last 40 years that was much greater than could be accounted for by population growth. When compared to national data, the same trend is found. Although not an ideal comparison as it is based on a slightly different age range and time frame, the Centers for Disease Control published a review of pediatric cancer incidence (birth to 20) for the years 2003-2014 by region and state and found Florida’s incidence of new cases to be 17.0 per 100,000 people over that entire time block (6). Looking past 2014 in the SPIRS database shows the incidence has increased even further in recent years. The simplest explanation is that children in Florida are truly getting cancer at a higher rate, irrespective of other factors, but it is most likely multifactorial. One reason is that there is improved reporting state-wide with the increase in the number of FAPTP centers during this time. In addition, more advanced imaging techniques lead to the discovery of more incidental benign tumors, which would be reportable and could skew the data. Previous analysis of this database in 2010 demonstrated cancer diagnosis clusters in certain southeast and northern parts of the state, which would make environmental causes more suspicious for this rising incidence (7). When examining the changing demographics of the state, another potential contributing factor comes to light: the most commonly diagnosed childhood cancer is acute lymphoblastic leukemia and this disease is roughly 30% more common in Hispanics than in non-Hispanic whites and nearly 2-fold more common in Hispanics than in African-Americans (8-10). These differences were noted in the populations of both Florida in 2000 (8) and California in 2016 (9) and their respective incidences are nearly identical. Compound this with the increasing number and proportion of Floridians that are Hispanic since 1980, we believe this factor cannot be understated. Yet another potential reason contributing to the increase in the number of oncology patients registered is that starting in approximately 2007, proton beam centers in Jacksonville and later Orlando began to receive referrals from outside the state, including international patients. In just the Jacksonville site alone, from 80-187 referrals were seen per year prior to a slow-down from the COVID19 pandemic (Dr. Scott Bradfield, personal communication, Sept. 8, 2021). One final factor that could be contributing to the increase in patients is increased referrals of new or relapsed children from Latin America and the Caribbean islands, many of whom come to Florida for treatment because of its proximity.
The significant increase in raw numbers and the percentage of these patients receiving care at FAPTP centers should equate to more standardized care with specialist expertise as well as greater access to clinical trials. Although the overall number of patients enrolled on large cooperative group trials increased, after a peak in clinical trial enrollment to 42% of all known patients through 1990, we have noted a steady decline since, down to 20% in the most recent 5-year block of data. So, while there has been improvement in access to specialized care for these patients, this has not translated into a higher rate of clinical trial enrollment, which merits further investigation and ongoing initiatives. The SPIRS data only accounts for CCG/POG/COG clinical trials, so a possible contributor to this trend could be patients enrolling on clinical trials outside of these large cooperative groups. Examples could include intra-institutional and pharmaceutical-led trials either at a FAPTP center or a larger referral cancer center outside the state of Florida after initial diagnosis. In such a case, the child would be enrolled on a clinical trial, but that data would not be captured by the SPIRS database.
Although the SPIRS data is not powered to allow us to know many of the reasons for patients not being enrolled, many reviews of barriers to clinical trial enrollment have been published, and the most common cause is usually found to be the lack of open upfront therapeutic trials (11-13). Some of that lack of availability may be because of the cooperative groups’ success in curing some of the most common diagnoses we see in pediatric oncology, such as standard risk acute lymphoblastic leukemia and Wilms tumor with what has become more standardized regimens over the years. This was not the case in the earlier years of the study period, and these reduced enrollments may simply reflect the successes of earlier efforts of the cooperative groups to establish better regimens.
And even though the average age of SPIRS patients have gone up over the last 40 years, it has not risen enough within the notoriously under-enrolled AYA range to explain the lower clinical trial enrollment by itself. Interestingly, Faulk et al looked at the SEER data and contemporaneous COG enrollment on therapeutic trials for children aged 0 to 19 years from 2004-2015 and saw a similar trend: only 19.9% of cancer patients enrolled on a COG therapeutic trial (14), down from 26.8% between 2000 and 2003 (11). In analyzing these data from 2000 to 2003, Lund et al found that the most underrepresented groups in COG trials were younger black and Hispanic children, Hispanic females and white teenagers aged 15-19 (11). Florida’s growing percentage of Hispanic children in the last 40 years may also be reflected in the enrollment data in this way. Health insurance status and the policies of specific carriers also may impact on clinical trial enrollment, as in recent years, some insurance carriers have been less enthusiastic covering patients enrolled on clinical trials, deeming it “experimental.”
Unfortunately, the SPIRS data does have other limitations in addition to not measuring clinical trial enrollment outside of the large cooperative groups. One is that the clinical trial enrollment data does not separate biological from therapeutic studies. The rates of enrollment between ages and racial/ethnic groups is also lacking, making it impossible to compare directly with Lund’s findings. There are undoubtedly disparities in access to healthcare in the United States which would include access to and enrollment on clinical trials, but unfortunately the SPIRS database does not have the appropriate data set to shed light on these disparities.
Another weakness is that outcomes are not tracked, so we do not know if greater access to specialized cancer care at FAPTP centers has led to improved overall and event-free survival in this patient population. There has been overall improved survival among pediatric cancer patients nationally over this time period (15), so it is very likely that the children of Florida have benefited along with other children across the country, but we do not have this specific outcome in our data set. It would also be interesting to have the racial breakdown of the patients seen at FAPTP centers in the database, to see if access to these centers was equal among black, white and Hispanic children in Florida. In this vein, St. Jude’s Children Research Hospital published their data from 2001 to 2007 and compared it to SEER data over the same time. They found that there were disparities in outcomes with some pediatric cancers such as neuroblastoma and acute myeloid leukemia between black and white children on the national level, but not in their own cohort, where all children were treated on the same regimens and insurance or lack thereof did not impact care (16). To have such data on a statewide level would add to the depth of our findings.
As the population of Florida has grown since 1980, so has the number of reported pediatric cancer patients in the state disproportionately along with the number of FAPTP centers. Our review of the unique SPIRS database, despite its limitations, shows that even though more children with cancer are being seen in these centers, there is still work to be done. Further mining of the data may allow us to better understand this population, but improving access to top quality cancer care across the state and addressing the racial, ethnic and age disparities, as well as dropping clinical trial enrollments would almost certainly lead to better outcomes, all of which should be measured prospectively moving forward.
Conflicts of Interest : The authors have none to declare.