The Unseen Financial and Human Toll of Clinical Trial Imaging Errors

The clinical trial landscape faces significant challenges impacting financial outcomes and patient well-being, with clinical trial imaging errors being among the most pressing issues. These errors can lead to substantial monetary losses and skewed trial results for trial sponsors. A recent publication highlights alarming error rates in clinical trial imaging data at major NCI-designated Comprehensive Cancer Centers, with error rates ranging up to 50% before implementing comprehensive clinical trials imaging informatics solutions ¹. Such inaccuracies can misrepresent treatment efficacy and delay drug approval processes by creating incorrect baseline measurements, discrepancies to the protocol in follow-up clinical trial imaging, and by the application of the wrong response criteria ². These inaccuracies skew critical endpoints like Progression-Free Survival (PFS) and Overall Response Rate (ORR), complicating the assessment of treatment efficacy.

For early phase and rare disease trials and for small to midsize biopharmaceutical companies, the high cost of imaging assessments utilizing Blinded Independent Central Reviews (BICRs) often leads to reliance solely on institutional image reviews, which, despite high-quality intentions, suffer from operational inefficiencies contributing to imaging errors. Even the most sophisticated NCI-designated Comprehensive Cancer Centers with dedicated imaging research cores struggle with these issues ^2, 3. Imaging errors have not been measured in less sophisticated environments, where they are likely worse. Additionally, CROs and large pharmaceutical sponsors who rely on BICRs to improve data quality also face challenges in operational efficiency and unnecessary costs, impacting trial timelines. Errors in clinical trial imaging assessments at the point of care propagate costs to CROs and sponsors as patients who were incorrectly included in trials must later be censored from the trial after high expense of enrollment, treatment, protocol violations, and trial delays. It is worth noting that comprehensive reviews, including BICRs, often occur at the end of the study rather than during live trials, which may contribute to these operational challenges. This highlights the need for comprehensive solutions to ensure data accuracy throughout the trial process, beginning at the point of care ².

Beyond the direct financial costs, clinical trial imaging errors affect the broader healthcare system. These include wasted funds from federal agencies like the FDA, NIH, and NCI, which often collaborate with and fund independent investigators and small biopharmaceutical companies to run early-phase trials. Additionally, there are incalculable traumatic impacts on patients who receive incorrect treatments or miss out on potentially life-saving therapies because of inaccurate determination of both initial and ongoing eligibility. Effective communication between sites, CROs, and sponsors, facilitated by a harmonized technology platform that enforces protocol compliance, is essential to mitigate these issues and enhance the overall efficiency and accuracy of clinical trials.

This article aims to quantify the costs associated with these errors and emphasize the urgent need for improved imaging accuracy in clinical trials starting from the point of care. We will explore the financial and systemic costs of clinical trial imaging errors, present data-driven estimates of their impact, and propose actionable solutions to enhance the accuracy and reliability of imaging data in clinical trials.

Clinical Trial Imaging Errors are Frequent and Costly

To understand the full impact of clinical trial imaging errors, examining the hard financial costs and the softer, more patient-centric impacts are essential. Imaging criteria compliance studies at three NCI-designated Comprehensive Cancer Center sites found 25%, 30%, and 50% errors in clinical trial imaging assessments, which were reduced to under 3% by implementing a comprehensive imaging informatics platform ². Such errors could lead to the inclusion of ineligible patients and the exclusion of eligible patients in trials, distorted data, misrepresentation of early efficacy signals, financial losses, and costly delays in drug approvals. In these studies, follow-up imaging response discrepancies from the trial criteria accounted for 29% of errors, skewing PFS and ORR endpoints, while missing or inaccurate measurements constituted 24% ².

These errors have far-reaching consequences, threatening trial integrity by not only misrepresenting treatment efficacy, but also increasing costs, extending trial duration, and ultimately delaying drug approvals. One of the most important successes for pharmaceutical companies is detecting a failed molecule early and stopping the program sooner. This early detection can provide significant cost savings, potentially even more than the acceleration of successful molecule approvals, due to the higher number of unsuccessful compounds and the high cost of continuing an ineffective trial. However, early phase trials that provide sponsors with the site data to make these early decisions are highly prone to the impact of clinical trial imaging assessment errors.². Such errors could lead to the inclusion of ineligible patients and the exclusion of eligible patients in trials, distorted data, misrepresentation of early efficacy signals, financial losses, and costly delays in drug approvals. In these studies, follow-up imaging response discrepancies from the trial criteria accounted for 29% of errors, skewing PFS and ORR endpoints, while missing or inaccurate measurements constituted 24% .

Additionally, one of the significant issues in clinical trials, particularly in oncology, is the lack of radiology point of care site reader training provided by sponsors. This gap might contribute to inconsistencies and errors in imaging data, which can undermine the integrity of trial results. As Jeffrey Sorenson, CEO of Yunu, points out, “The reality is that sponsors do not conduct adequate, protocol-specific training for radiologists performing site assessments, leaving them to navigate complex imaging assessments without proper guidance. This oversight, along with broken workflows and inadequate measurement tools are just a few of the many critical factors contributing to the high error rates we see in clinical trial imaging.“

Addressing these issues is crucial for improving clinical trial efficiency and imaging data quality. Enhancing the accuracy of imaging data ensures that the right patients are enrolled, resources are not wasted on ineffective treatments, and the timeline for bringing effective treatments to market is optimized.

Cost Analysis

In a previous article, we demonstrated imaging errors at the point of care in clinical trials ranged from 25% to 50% ¹, contributing to a 10% patient censor rate upon secondary central review ^1, 4, 5 (which has been further verified by Yunu’s imaging database, yielding a similar censor rate). These errors, which result in the censored patients’ data being excluded from the trial, often stem from systemic issues, such as not following the protocol correctly rather than inter-reader variability. Repeating the same read under the same suboptimal conditions would likely result in similar error rates. These inaccuracies can lead to the inclusion and inappropriate treatment of ineligible patients, impacting data quality on imaging that has already been done. Even though BICRs are conducted on a subset of trials (most trials do not have the budget to pay the high costs of BICR) with a careful eye toward protocol compliance and reader training, they are performed off-site and later in the trial process. Therefore, the initial 25% to 50% error rate still affects the trial as it influences patient eligibility and ongoing participation as seen in the high 10% censor rate ^1, 4, 5. These errors result in wasted resources on treatments and follow-ups that yield no valuable data, prolonging trial durations and escalating costs ⁶.

This section will help trial stakeholders quantify study budgets as if there were no imaging errors and a new imaging informatics platform were used programmatically across all sites in a trial. This new algorithm models imaging data quality issues during the trial, necessitating redoing visits or enrolling new patients. With 25% to 50% of time points having errors as read at the point of care, the cost of these scenarios is consistently incurred. If errors are eliminated from the outset, the algorithm estimates reduced study costs and duration, thus avoiding re-enrollment and repeated visits. Additionally, the algorithm includes adjustments and sensitivity analyses to mimic real-world settings, such as added project management costs and delays in enrollment.

Assumptions in the model: The median cost per patient in oncology clinical trials is $100,000 ⁷. The mean enrollment is 46.9 patients in Phase 1 and 214.8 in Phase 2 oncology trials, and the mean number of follow-up days is 56.4 for Phase 1 and 134.9 for Phase 2, with total trial durations averaging 640 days and 1,027 days, respectively ⁸. The mean actual cost for a phase 1 oncology trial is $14.2 million, and $25.5 million for a phase 2 oncology trial ⁹. These benchmarks are used to help calculate the financial impact. The mean direct cost of conducting a Phase 2 trial is estimated at $23,737 per day ¹⁰.

By applying the assumptions for a phase 1 trial with a 50% imaging error rate, sponsors can save $1.1 million, representing an 8% cost reduction, and shorten the trial duration by 57 days, equating to a 9% time savings. In a phase 2 trial with the same error rate, the savings could be nearly $1.8 million, a 7% cost reduction, and the trial could be completed 192 days faster, yielding a 19% time savings.

Impacts of Inaccurate Site Imaging Data

Inaccurate imaging data in clinical trials can have profound implications, affecting patient cohorts and the accuracy of trial results. Regulatory bodies and sponsors are increasingly focusing on the need for precise imaging data to ensure reliable trial outcomes. Discrepancies in imaging assessments can lead to incorrect conclusions about treatment efficacy, skewing critical endpoints such as PFS and ORR. The result is prolonged trial durations, increased costs, and delayed access to potentially life-altering treatments for patients. In actuality, sponsors discovering and correcting imaging errors mid-trial is much less likely, as many sponsors in early-phase trials rely on institutional imaging readers without BICRs. If BICRs are performed, they often occur after the trial concludes, which has been shown to cause the aforementioned 10% censor rate ^1, 4, 5.

An alternative costly scenario occurs when, rather than early-phase imaging errors resulting in incorrect drug discontinuation, medical products are erroneously advanced into later phases based on false positive signals from imaging errors. These errors may lead to later-phase trials, where more rigorous and expensive imaging data quality methods and BICRs are implemented, ultimately revealing the true inefficacy of the product. This results in a costly failure at a later stage which is not accounted for in the calculator provided, which would dramatically increase the potential cost savings of accurate, reliable site reads. In rare cases, a medical product might be taken off the market entirely due to imaging inaccuracies, as demonstrated in the case study below. Although early-phase trials alone do not lead to approved drugs, imaging errors at this stage can still cause significant delays and financial losses as issues are identified and corrected in later phases.

Call to Action: Leveraging Technology to Enhance Imaging Accuracy in Clinical Trials

Small biopharmaceutical companies often face the dilemma of balancing tight budgets with the need for high-quality imaging data in early-phase trials. The cost of BICR can be prohibitive, leading many to rely on institutional site reads without rigorous reviews or adequate imaging assessments on protocol-compliant informatics platforms, resulting in high imaging error rates ². Research institutions strive to review images effectively, but they operate under challenging circumstances that almost guarantee failure without a comprehensive clinical trial imaging informatics solution, even at the largest, most well-resourced institutions. These challenges include discrepancies in follow-up imaging response from trial-specific rules, incorrect baseline measurements, and application of incorrect response criteria, all of which contribute to significant imaging errors.

The benefits of adopting advanced clinical trial imaging technology are equally compelling for imaging CROs and large pharmaceutical sponsors. With the appropriate comprehensive imaging informatics technology, imaging CROs can eliminate error-prone and time-intensive manual data uploads and streamline workflows, ensuring pristine data and unparalleled radiology reader collaboration across all sites, trials, and stakeholders. This accelerates trial timelines and ensures audit readiness and compliance with regulatory standards. Large pharma sponsors, on the other hand, can achieve faster turnaround times, gaining earlier insights and making quicker go/no-go decisions with real-time access to high-value trial imaging data. Additionally, both enterprises can benefit from efficiently using imaging technology to implement and document radiology reader training, ultimately improving imaging data quality. This optimization of resource allocation and trial efficiency turns clinical trial imaging risk into ROI, ensuring that imaging becomes an advantage rather than a hurdle, supporting the delivery of innovative therapies to market faster and more reliably.

Even when BICR is employed, the inherent delays can impact trials substantively. By leveraging advanced imaging informatics platforms and enforcing stringent quality and compliance process and standards, small biopharma sponsors and CROs can ensure high data quality without the hefty costs associated with BICRs. This allows BICRs to be reserved for later-phase trials when more funding is available while not sacrificing the reliability and efficacy of early-phase trials. Implementing these technologies can assure trial compliance by providing accurate signals and maintaining data integrity from the outset, thus avoiding costly re-enrollment or censoring of patients and repeated study visits or patient dropout due to imaging errors and delays. Reliable point of care imaging assessments also avoid erroneously advancing ineffective therapies on early-phase trials that should be terminated.

The tide is turning. Currently, over 20% of the NCI-designated Comprehensive Cancer Centers in the US have implemented a new platform by which they manage and monitor all clinical trial imaging assessments with rigorous yet highly efficient methods ¹⁶. Additionally, good workflow results in clean data that can be processed with AI models to extract additional insights, enhance data value, and lead to better-informed trial designs and increased efficiency.

Conclusion

Addressing clinical trial imaging errors in clinical trials is imperative for ensuring accurate and reliable data, directly impacting treatment efficacy, trial integrity, and patient outcomes. By leveraging advanced imaging informatics platforms and automation and insight extraction tools, sites where patients are enrolled, small biopharmaceutical companies, CROs, and large pharmaceutical sponsors can all benefit by significantly enhancing data quality, reducing costs, and accelerating trial timelines. Implementing these technologies mitigates the risk of costly errors and delays leading to patient dropout, re-enrollments and repeated study visits while fostering more efficient and effective clinical trials. Ultimately, adopting these innovative solutions will support the timely delivery of life-saving therapies to patients and uphold the integrity of the clinical research process.

This article is sponsored by Yunu.

1. Cruz A, Lankhorst B, McDaniels H, Weihe E, Correa E, Nacamuli D, Somarouthu B, Harris GJ. The complete workflow solution for quantitative imaging assessment of tumor response for oncology clinical trials. Presented at AACI-CRI Conference, Chicago, IL, 2024. ↩︎
2. Urban T, Ziegler E, Leary M, Somarouthu B, Correa E, Basinsky G, Nacamuli D, Sadow CA, O’Malley R, Wang C, Van den Abbeele AD, Harris GJ. Precision Imaging Metrics: Changing the way clinical trial imaging assessment is managed. Presented at AACI-CRI Conference, Chicago, IL, 2018. ↩︎
3. https://www.clinicaltrialvanguard.com/article/urgent-warning-dangers-and-costs-of-clinical-trial-imaging-errors/ ↩︎
4. https://www.clinicaltrialvanguard.com/wp-content/uploads/2024/05/080411.pdf ↩︎
5. https://grants.nih.gov/grants/guide/pa-files/PAR-17-167.html ↩︎
6. https://ascopubs.org/doi/pdfdirect/10.1200/JCO.2022.40.16_suppl.e13603 ↩︎
7. https://www.statista.com/statistics/1197095/clinical-trial-cost-per-patient-by-therapy-area/ ↩︎
8. Getz, K., Smith, Z. & Kravet, M. Protocol Design and Performance Benchmarks by Phase and by Oncology and Rare Disease Subgroups. Ther Innov Regul Sci 57, 49–56 (2023). https://doi.org/10.1007/s43441-022-00438-5 ↩︎
9. Tufts Center for the Study of Drug Development. “Cost variation and mis-estimation characterize clinical trial budgets, particularly in early phases.” Tufts CSDD Impact Report, vol. 24, no. 2, March/April 2022. Zak Smith and Ken Getz. ↩︎
10. https://www.appliedclinicaltrialsonline.com/view/how-much-does-a-day-of-delay-in-a-clinical-trial-really-cost- ↩︎
11. https://www.clinicaltrialvanguard.com/wp-content/uploads/2024/05/080411.pdf ↩︎
12. https://grants.nih.gov/grants/guide/pa-files/PAR-17-167.html ↩︎
13. https://www.financialmirror.com/2011/06/29/roche-seeks-more-time-for-avastin-in-breast-cancer/ ↩︎
14. https://trialsjournal.biomedcentral.com/articles/10.1186/s13063-017-1870-2 ↩︎
15. https://www.appliedclinicaltrialsonline.com/view/role-independent-review-oncology-trials ↩︎
16. http://dx.doi.org/10.1117/12.3013791 ↩︎

 

Moe Alsumidaie

Website | Moe Alsumidaie is Chief Editor of The Clinical Trial Vanguard. Moe holds decades of experience in the clinical trials industry. Moe also serves as Head of Research at CliniBiz and Chief Data Scientist at Annex Clinical Corporation.