Current trends and outcomes in cellular therapy activity in the United States, including prospective Patient Reported Outcomes data collection within the CIBMTR registry

Abstract

Background:

Center for International Blood and Marrow Transplant Research (CIBMTR) prepares an annual set of summary slides to detail the trends in transplantation and cellular therapies. For the first time in the 2023 summary slides, CIBMTR incorporates data for patients receiving chimeric antigen receptor T-cell (CAR-T) infusions. In addition, the data on patient-reported outcomes (PROs) is also included.

Objectives:

This report aims to update the annual trends in US HCT activity and incorporate data on the use of CAR-T therapies. Here we also aim to present and describe the development, implementation, and current status of the PRO data collection.

Study Design:

In August 2020, CIBMTR launched the Protocol for Collection of Patient Reported Outcomes Data (CIBMTR PRO Protocol). The CIBMTR PRO Protocol operates under a centralized infrastructure to reduce burden to centers. Specifically, PRO data is collected from a prospective convenience sample of adult HCT and CAR-T patients who received treatment at contributing centers and consented for research. Data are merged and stored with the clinical data and used under the governance of the CIBMTR Research Database Protocol. Participants answer a series of surveys developed by the Patient Reported Outcomes Measurement Information System^© (PROMIS) focusing on physical, social and emotional, and others measures assessing financial well-being, occupational functioning, and social determinants of health. To complement traditionally measured clinical outcomes, the surveys are administered at the same timepoints that clinical data is routinely collected.

Results:

As of September 2023, PRO data from 993 patients across 25 different centers has been collected. With the goal of incorporating these important patient perspectives into standard clinical care, CIBMTR has added the PRO data to Data Back to Centers (DBtC). Through expanding the data types represented in the registry, CIBMTR aims to support holistic research accounting for the patient perspective in improving patient outcomes.

Conclusion:

PRO data at CIBMTR aims to provide the foundation for future large scale, population-level evaluations to determine areas for improvement, emerging disparities in access and health outcomes (eg, by age, race, and ethnicity), and new therapies that may impact current treatment guidelines. Continuing to collect and grow the PRO data is critical for understanding these changes and identifying methods for improving patient quality of life.

INTRODUCTION:

As the use of hematopoietic cell transplantation (HCT) and adoptive cellular therapies evolves and expands, it is important to monitor trends in activity and outcomes. The Center for International Blood and Marrow Transplant Research (CIBMTR) collects longitudinal clinical data for recipients of HCT and other cellular therapies like chimeric antigen receptor T-cell (CAR-T) and gene therapies. As the contract holder for the Stem Cell Therapeutic Outcomes Database (SCTOD) since 2005, CIBMTR has collected data on all allogeneic HCTs performed in the United States as mandated center reporting, providing a true denominator of all U.S. allogeneic HCT activity. CIBMTR also collects information submitted on a voluntary basis for about 85% of all autologous HCTs in the U.S. and 60–70% of CAR-T therapies, beginning in July 2016. As of September of 2023, CIBMTR has information from 253,127 first autologous HCT recipients, 183,327 first allogeneic HCT recipients, and 10,347 first CAR-T recipients. This extensive data captures patient, disease, and treatment characteristics, as well as exposures, clinical outcomes, and late effects. The CIBMTR registry serves as a platform for registry and statistical experts to engage with the cellular therapy community and address important questions. Additionally, it provides current and relevant information for clinical investigators.

Patient reported outcomes (PROs) are reports by patients of their symptoms and functioning without interpretation by others, such as a clinician (1). Prior studies confirm PROs provide nuanced information on patients’ symptoms and quality of life (QOL) beyond what clinical data can provide, particularly regarding the less observable symptoms (2,3). Recognizing the growing importance of the patients voice through PROs in assessing and predicting patient QOL related to cellular therapies, CIBMTR developed the Protocol for Collection of Patient Reported Outcomes Data (hereafter CIBMTR PRO Protocol) to supplement traditional clinical outcomes data. The PRO data within the CIBMTR registry serves as an additional source of information to create hypothesis driven analyses that, importantly, include firsthand accounts of patient experiences and perspectives.

This report updates the annual trends in US HCT activity and incorporates data on the use of CAR-T therapies for the first time (full slide set posted on https://cibmtr.org/Files/Summary-Slides--Reports/Summary-Slides-2023-website-version.pptx [retrieved 06/18/2024]). Alongside the clinical data, CIBMTR has provided a set of slides focused on PROs collected in the registry, and here we present and describe the development, implementation, and current status of the PRO data collection.

MATERIALS AND METHODS:

Electronic PRO (ePRO) Development.

CIBMTR began building the ePRO infrastructure in 2017 by first convening a transdisciplinary advisory group made up of HCT physicians, PRO experts, patients, and caregivers with the goal of providing recommendations for the measurement approach and timepoints for the CIBMTR PRO data collection. The group recommended nine domains covering physical, mental, and social well-being from the Patient Reported Outcomes Measurement Information System© (PROMIS)(4) and timepoints aligned with the clinical data collection at CIBMTR (Figure 1). Additionally, measures and items were recommended to assess financial toxicity, occupational functioning, and social determinants of health.

Figure 1.

Patient Reported Outcomes Collected on the CIBMTR PRO Protocol Survey

Centralized data collection.

Next, the CIBMTR developed the ePRO infrastructure. The goal of the infrastructure was to centralize all aspects of the research process (e.g., recruitment, consent, data collection, patient follow up) and link to the SCTOD. To accomplish this, Salesforce, a customer relationship management system, was implemented to track patient outreach and schedule research tasks across research team members. Qualtrics was used to build the PRO surveys. A benefit to using the PROMIS measurement system is the ability to utilize Computer Adaptive Test (CAT) versions of the measures, which leverage a patient’s prior responses to a question to inform future items. As a result, patients complete as few items as possible without losing precision in the scores. The PROMIS API is directly connected to the Qualtrics system to allow for CAT administration and scoring. PRO data are stored in the CIBMTR Research database with unique CIBMTR identifiers that directly link the PROs to the patient-level clinical data. Use of the data is governed by the CIBMTR Research Database Protocol providing investigators broader use of all data available through CIBMTR (Figure 2).

Figure 2.

CIBMTR Patient Reported Outcomes Survey Workflow Process and Data Storage Journey

The centralized ePRO infrastructure was built to seamlessly integrate with existing CIMBTR processes which allows for minimal effort at the centers. The CIBMTR PRO Protocol is governed under a centralized IRB at NMDP, which alleviates any IRB requirements at centers. The only task for centers is already part of the CIBMTR consent workflow: centers complete the CIBMTR Research Database Protocol consent form to allow patient data to be sent to CIBMTR. Part of that consent asks patients if they agree to be contacted for future research. If patients respond “yes” to this question, the center provides the patients contact information to CIBMTR, the centralized infrastructure described above is triggered, and CIBMTR begins outreach to eligible patients. Currently, eligible patients are any adult patient who receives a first cellular therapy at a center who agrees to enroll patients, is able to comfortably read English or Spanish, and has agreed to be contacted for future research. If a patient enrolled in the PRO protocol receives a second therapy (HCT or other cellular therapy), they continue to complete timepoints based on their first treatment.

Piloting of the system took place in a cross-sectional observational study of patients age ≥55 who underwent allogeneic HCT for myelodysplastic syndromes (MDS) with the goal of understanding whether significant differences would exist between patients who do and do not complete PROs(5). Results confirmed no significant differences between patients who do and do not participate in PROs. Barriers to enrollment were identified (e.g., unable to contact or implicit refusal) and CIBMTR was able to implement strategies to address these barriers.

Initial Roll out of CIBMTR PRO Protocol.

After determining the feasibility of the ePRO infrastructure as described above, a phased approach to enrollment onto the observational prospective longitudinal PRO protocol began in 2020 (Figure 3). Three champion centers rolled out recruitment to allow for logistics testing and alignment, with the first patient enrolled onto the CIBMTR PRO protocol in August of 2020. Within a few months, enrollment of alloHCT and autoHCT recipients expanded to seven centers. The gradual phased approach was intended to confirm workflows and processes related to the centralized ePRO infrastructure and allow the team time to troubleshoot any issues that arose. In September 2021, additional interested sites were invited to join with accompanying informational webinars. To avoid patient burden, centers are asked to opt in to the CIBMTR PRO Protocol to ensure the data collection is not duplicative of efforts at the center. The first CAR-T patient was enrolled in March 2021, the 500^th patient in August 2022, and the first Spanish-speaking patient in November 2022.

Figure 3.

Timeline of the CIBMTR electronic Patient Reported Outcomes (ePRO) Infrastructure

Data Back to Centers (DBtC).

CIBMTR is committed to returning data to the centers for their own purposes, with added visualizations where possible. Through the CIBMTR DBtC portal, centers receive scores for all patients who complete surveys at their center and have consented to us sharing their data with their care team. Data is aggregated into line graphs and available within the DBtC dashboard where clinicians can filter by clinical information (e.g., infusion type). PRO data are downloadable and updated monthly.

Statistical Analysis PRO Data.

Descriptive statistics on patient, infusion, disease, and sociodemographic characteristics were performed for two groups of patients: (1) first, all patients who enrolled into the CIBMTR PRO Protocol and (2) second, the subset of patients who have completed a PRO survey prior to infusion. Where patient sex and race/ethnicity was completed on the PRO survey, that data was used. If the data was missing on the PRO Protocol, the clinical data as submitted by the centers was used. Age, sex, race, and ethnicity are also collected on the standard CIBMTR data collection forms.

Statistical Analysis Clinical Outcomes Data.

Data reported to CIBMTR are used to estimate total HCT and CAR-T numbers. Overall survival probabilities are presented according to disease, disease status, donor type, recipient age, and conditioning regimen intensity. Causes of death are reported. Trends in the total numbers of transplantations performed in the United States up to 2022 are analyzed. Survival analyses use Kaplan-Meier estimators at the 3-year time point with 95% confidence intervals (CIs). Estimates of survival curves and comparisons across curves are univariate and do not adjust for potentially relevant contributing factors.

RESULTS:

As of September 2023, the CIBMTR PRO Protocol has enrolled 993 patients from 25 centers, with an enrollment rate of 34%. Overall, patients who enroll in the CIBMTR PRO Protocol are similar to those who do not enroll, with those enrolled slightly older (average 62 yr old versus 60) and more autoHCT patients enroll (53% versus 50%) (data not shown). Of enrolled patients, 202 (20%) completed baseline surveys, 56 Day 30 surveys, 737 Day 100 surveys, 767 Day 180 surveys, 529 1 Year surveys, 131 Year 2 surveys, and 3 Year 3 surveys completed (Figure 4). Considering the full patient population in the current convenience sample collected, most (39%) were diagnosed with a plasma cell disorder (PCD), followed by leukemias and lymphomas (22% and 21%, respectively). 53% of the patients enrolled into the CIBMTR PRO protocol received an autologous HCT, 36% received an allogeneic HCT, and 11% received CAR-T. Patient baseline Karnofsky scores vary with 53% of patients with a score of 90 or above and 42% with scores less than 90. Majority of patients are male (n=537; 54%). Over three-fourths of patients are age 50 or older and 75% are non-Hispanic White. 6% of patient are Hispanic, White and 6% are non-Hispanic Black or African American. The full set of summary slides for PROs can be viewed at https://cibmtr.org/Files/SummarySlides--Reports/CIBMTR_PROSummary-Slides_2023-website-version.pptx (retrieved06/17/2024).

Figure 4.

Bar Chart Illustrating the Number of PRO Surveys Completed at each Timepoint of Data Collection

Patients with Baseline PRO Data in CIBMTR PRO Protocol.

Of the patients enrolled into the CIBMTR PRO Protocol, 20% (n=202) completed a pre-infusion timepoint survey. These patients overall have similar clinical and patient characteristics compared to the overall sample, with some differences (Table 1). Among those with a pre-infusion timepoint, there are fewer patients treated for lymphomas and PCD and more patients with leukemias (n=51; 25%) compared to the full sample. Compared to the full sample, there is a larger proportion of allogeneic HCT (n=88; 44% in pre-infusion patients compared to 36% in the full sample) for patients with pre-infusion data. KPS scores are similar to the full sample. Over half of patients with pre-infusion surveys are female (n=110; 55%) and there is a larger proportion of patients under the age of 30 (16% with pre-infusion surveys as opposed to 7% in the full sample). The racial and ethnic diversity is lower in those with pre-infusion surveys (87% of the patients are non-Hispanic white) compared to the full sample (Table 1).

Table 1.

Clinical and Patient Characteristics of Patients enrolled into the CIBMTR PRO Protocol by the full sample and those who have a pre-infusion survey completed

	Total Sample		With Pre-infusion
	N	%	N	%
Number of patients	993		202	20%
Number of centers	25		13
Clinical Characteristics
Primary indication for infusion
Leukemias	219	22%	51	25%
Myelodysplastic Syndrome	86	9%	20	10%
Lymphomas	206	21%	32	16%
Plasma cell disorders	384	39%	52	26%
Other	98	10%	47	23%
Graft type
Bone marrow	36	4%	10	5%
Peripheral blood	809	81%	149	74%
Umbilical cord blood	8	1%	1	1%
Other cellular therapy	110	11%	18	9%
Not reported	30	3%	24	12%
HCT-CI
0	246	25%	30	15%
1	166	17%	31	15%
2	150	15%	32	16%
3+	380	38%	77	38%
Undetermined	6	1%	2	1%
Not reported	45	5%	30	15%
HCT-CI for patients receiving HCT
0	211	24%	23	13%
1	150	17%	31	17%
2	139	16%	31	17%
3+	348	40%	74	41%
Undetermined	5	1%	1	1%
Not reported	26	3%	20	11%
HCT-CT for patients receiving CT
0	35	32%	7	39%
1	16	15%	1	6%
2	11	10%	3	17%
3+	32	29%	1	6%
Undetermined	1	1%	6	33%
Not reported	15	14%	6	33%
Infusion type
Allogeneic HCT	354	36%	88	44%
Autologous HCT	525	53%	92	46%
CAR-T	110	11%	18	9%
Not reported	4	<1%	4	2%
Year of infusion
2020	15	2%	14	7%
2021	244	25%	37	18%
2022	559	56%	67	33%
2023	171	17%	80	40%
Missing	4	0%	4	2%
KPS pre-infusion
90–100	529	53%	107	53%
< 90	419	42%	66	33%
Not reported	45	5%	29	14%
Patient characteristics
Sex
Male	537	54%	67	33%
Female	417	42%	110	55%
Something else	22	2%	22	11%
Not reported	17	2%	3	2%
Age
18	43	4%	29	14%
18–29	31	3%	4	2%
30–39	54	5%	12	6%
40–49	90	9%	21	10%
50–64	386	39%	59	29%
65+	389	39%	77	38%
Race and ethnicity
White	744	75%	176	87%
White, Hispanic or Latino	55	6%	9	4%
Black or African-American	64	6%	6	3%
Black or African-American, Hispanic or Latino	1	<1%	0	0%
Asian	47	5%	6	3%
Asian, Hispanic or Latino	1	<1%	0	0%
American Indian or Alaska Native	3	<1%	0	0%
Native Hawaiian or Pacific Islander	2	<1%	0	0%
Hispanic or Latino, Race NOS	33	3%	2	1%
Not reported	43	4%	3	1%
Race
White	799	80%	185	92%
Black or African-American	65	7%	6	3%
Asian	48	5%	6	3%
American Indian or Alaska Native	3	<1%	0	0%
Native Hawaiian or Pacific Islander	2	<1%	0	0%
Not reported	76	8%	5	3%
Ethnicity
Hispanic or Latino	94	9	15	7%
Not Hispanic or Latino	864	87	184	91%
Not reported	35	4	3	2%
Household Annual Income (self-reported)
Less than $20,000	12	2%	7	3%
$20,000 - $49,999	39	4%	37	18%
$50,000 - $79,999	42	4%	38	19%
$80,000 - $99,999	30	3%	27	13%
$100,000 - $199,999	43	5%	40	19%
$200,000 or more	18	2%	16	8%
Don’t know	16	2%	14	7%
Not reported*	794	80%	23	11%
Education
High school diploma or equivalent (GED)	22	2%	35	17%
9th - 12th grade, no diploma	6	1%	3	1%
Some college, no degree	39	4%	35	17%
Vocational or Associate’s degree	30	3%	28	14%
College degree (B.A./B.S.)	53	5%	50	25%
Advanced degree (e.g., Master’s or Doctorate program)	63	6%	59	29%
Don’t know	1	<1%	0	0%
Not answered	1	<1%	0	0%
Missing*	778	78%	10	5%

^*Household Income and Education, along with our other sociodemographic questions were only asked at baseline and 1 year timepoints until 11/17/2023. Since we were only able to capture 20% of patients at baseline, this led to a significant number of missing sociodemographic information. On 11/17/2023, we implemented new logic which asked patients sociodemographic questions at the first survey they completed, regardless of the timepoint at which they started to improve our missingness on sociodemographic questions.

An example of the type of data and trends described from this data: Physical Function Trajectory.

Figure 5 displays the average physical function score at each timepoint for patients with available at that timepoint. The average patient pre-infusion reported a score of 42.1, which translates to mild impairment compared to the national average of the US general population. The average physical function scores fall within the range of mild impairment at day 100 and day 180. At one-year post-infusion the average physical function score (45.9) is above the U.S. Cancer norm and falls within the normal limits of functioning compared to the U.S. general population.

Figure 5.

Trajectory of Average Physical Function at Pre-infusion and Four timepoints post-infusion among patients who enrolled in the CIBMTR PRO Protocol (n=993)

Footnote: The PROMIS physical function domain was scored using response pattern scoring through the API system linked to Qualtrics. PROMIS domains scores leverage the T-score system in which a score of 50 reflects the average of the US population and 10 points characterize 1 standard deviation within the US population. Clinically meaningful changes in scores are disease and domain specific and range from 2- to 5-points(12,13). Therefore, we considered 5 points, or ½ a standard deviation, to be a meaningful change.

DISCUSSION:

The CIBMTR 2022 annual activity report highlights an increase in the number of total HCTs and CAR-T performed each year, with continued improvements in survival in the recent era (HCT in 2016–2021) (https://cibmtr.org/CIBMTR/Resources/Summary-Slides-Reports). The number of older, ethnically diverse, and patients with other co-morbidities receiving cellular therapy continue to increase. Patients in these categories often have increased supportive care needs and understanding their holistic experience can be improved through collection and analysis of PROs(6–8).

With growing recognition of the importance of PROs to understand patient experience within HCT research, the CIBMTR provides a critical source of such data. While most PRO studies in the HCT space report data from a single site, have small numbers, or collect on prospective clinical trials, the CIBMTR PRO Protocol provides large-scale, real world data that can be directly linked to clinical data, and is collected in a cost effective setting. This unique data provides clear utility in several ways, including potential benchmarks for patient well-being(9) within the HCT/CT population and the ability to compare across other diseases and treatment populations.

Since the first patient was enrolled in August 2020, the CIBMTR PRO Protocol has systematically and centrally collected information in a broad range of functional, symptom and socio-demographic domains. The data described here illustrates the expanding data, particularly patients with pre-infusion data, that are available to the research community to answer patient centered outcomes research questions. The trajectory of physical function scores, as one example, showing improvement from pre-infusion through 2 years post-infusion highlight patients physical function scores return to U.S. cancer and general population scores by 1 year following treatment. This confirms previous literature in observational(5) and clinical trial(10) studies of patients receiving an allogeneic HCT.

As with other registry PRO data collection efforts, the CIBMTR PRO Protocol recognizes response rates and selection bias as limitations to the generalizability of the data. The enrollment rates at the CIBMTR PRO Protocol are lower than other published registry enrollmend rates, though importantly other registries collecting PROs alongside clinical data are among different patient populations with different treatment and recovery trajectories (e.g., 75% of the studies assessed in a 2020 systematic review by Wang and colleagues were among orthopedic surgery patients) (11). More broadly, registry response rates for PROs have declined since the 1990s. The initial CIBMTR PRO Protocol ramp up sought to prioritize ensuring sustainable processes, determining logistics, and expanding the visibility of the mechanism, rather than response rates or representativeness of sample, however, these are important next phases of the registry effort. Though response rates provide a useful metric to understand generalizability, there are alternative attributes to a sample to understand how representative it is of the population of interest.

The CIBMTR PRO protocol data currently remains a convenience sample of patients from centers who agree to allow CIBMTR to contact to their patients, which introduces selection bias. This is common among PRO registry data collection (11). As the protocol expanded, improvements to the process and survey content were made to balance the essential information while not overburdening patients. Therefore, some patients may have more complete data on sociodemographic characteristics (e.g., household income and education) than others. To enhance the research potential of the data, CIBMTR has put strategic efforts in place. First, plans to improve pre-infusion response rates include identifying and targeting centers with the infrastructure to provide patient contact details early enough to initiate contact prior to pre-infusion, and implementing improved patient engagement strategies. Pre-infusion data critically provide a point of comparison for the health and quality of life impacts of HCT and CAR-T treatments. The number of new patients enrolled decreased in 2023 due to an initiative to focus resources on capturing patients pre-infusion rather than enrolling any patient at any timepoint. Second, a patient engagement focused group of CIBMTR team members, consulting with the Consumer Advocacy Committee, a group of patients and patient advocates, have revisited recruitment process and materials to ensure they are patient friendly with hopes of improving response rates. Third, expansion to pediatric PRO collection presents another opportunity for growth of the CIBMTR PRO Protocol and representation of additional age groups as the current eligibility criteria is for patients >18 years old, most of whom (78%) are over the age of 50.

To promote and improve access to PRO data, the CIBMTR launched the PRO portion of the DBtC platform in 2022. The goal of the DBtC platform is to allow care teams to visualize and use PRO data and accelerate the process of implementing PROs into clinical care. Care team members can access line graphs similar to Figure 5 of their patients who completed PROs within DBtC. Upcoming enhancements to DBtC PROs are planned such as adding the social determinants of health characteristics, including financial toxicity and developing methods for care team members to download PRO data in a more user friendly format. More regular updates of the PRO data are also planned to potentially allow for more practical programmatic or clinical use.

PRO data for research studies are available to the community through existing systems such as the working committees. Researchers are encouraged to propose studies to the disease or research content specific CIBMTR Working Committee and indicate their study will use PROs on their Working Committee Proposal. CIBMTR also supports grant applications or funded studies which could include PRO data already collected, or additional questions of interest.

CONCLUSION:

For over 50 years, CIBMTR’s data have described constantly evolving conditioning regimens, donor types/access, and improvements in supportive measures. As the use of HCT continues and the prevalence of adoptive cellular therapies therapies grows, constant reevaluation of treatment and patient outcome trends is necessary. CIBMTR has expanded to include data from recipients of these adoptive cellular therapies, such as CAR-T. The PRO data will provide the foundation for future large scale, population-level evaluations to determine areas for improvement, emerging disparities in access and health outcomes (eg, by age, race, and ethnicity), and new therapies that may impact current treatment guidelines. A continued focus on PRO data collection is critical for understanding these changes and identifying methods for improving patient quality of life.

CIBMTR is committed to collecting high quality data and making it accessible for those who work to improve outcomes in recipients of HCT and adoptive cellular therapies. This report, along with the summary slides posted at https://www.cibmtr.org provide an updated and accurate data resource for clinicians, patients, and investigators.

Highlights.

• Continued improvements in survival for patients receiving cellular therapies

• Patient reported outcomes bring patient perspective to clinical registry data

• Cellular therapy registry has patient reported data on about 1000 patients

• Patient reported outcomes data is shared back to participating centers