By Bill Atkinson 

 What it all came down to was human error.  

 ZNA Cadix hospital in Antwerp, Belgium last week took the unusual step of firing its entire team of radiologists – seven in all.  

 Why? Too many errors. Too many incomplete and inaccurate diagnoses. 

 ZNA, the hospital’s parent, said radiologists from its other campuses would fill in the gaps at ZNA Cadix while an internal investigation is underway. 

 “For us, the safety of the patients is the utmost importance,” said Dr. Katrien Bervoets, ZNA’s chief physician. “That is why we did not hesitate to intervene immediately. We do not want to and cannot wait for the final report of the medical audit and additional research. We will investigate this thoroughly and of course notify the patients involved.” 

 Radiologists have long been under pressure especially as the medical industry looks to artificial intelligence to improve performance, speed, accuracy and reduce costs.  

 There is little doubt that the ZNA Cadix debacle will fuel a deeper discussion on why hospitals should investigate ways to improve performance using technology. Radiology has often been the place where hospitals can wring out efficiencies and make improvements.  

 The threat of AI putting radiologists out of work has for years hung over the profession. 

 “As deep learning algorithms and narrow A.I. started to buzz especially around the field of medical imaging, many radiologists went into panic mode,” writes Dr. Bertalan Mesko, in The Medical Futurist. 

 But Mesko notes that there is “a lot of hype and plenty of fear around artificial intelligence and its impact on the future of healthcare. There are many signs pointing toward the fact that A.I. will completely move the world of medicine.” 

 Yet, like others who follow the industry, Mesko believes that radiologists who adopt AI will replace those who don’t use it. 

“Radiologists’ creative work will be necessary in the future to solve complex issues and supervising diagnostic processes; but AI will definitely become part of their daily routine in diagnosing simpler cases and taking over repetitive tasks,” Mesko writes. 

 Yet, mistakes are a fact of life in hospitals and “diagnostic errors are still the most commonly reported medical mistakes, affecting  an estimated 12 million adults in outpatient care each year,” according to the Wall Street Journal’s “Can AI Help Doctors Come Up With Better Diagnoses?” 

 A study, “Prevalence of harmful diagnostic errors in hospitalised adults: a systematic review and meta-analysis” concludes that in the U.S. there are nearly 250,000 “harmful diagnostic errors” each year. 

“Of the 136 diagnostic errors that were described in detail, a wide range of diseases were missed, the most common being malignancy (n=15, 11%) and pulmonary embolism (n=13, 9.6%),” according to the study. 

Dr. Gabriel Krestin, an expert in radiology who was also the acting chairman of the Department of Radiology at University Hospital in Zurich, Switzerland, said in an interview with FLUIDDA magazine that current image interpretation in radiology leaves room for inconsistencies.  

 “Images are actually gray shadows, and it is an art to translate these shadows into meaningful information,” said Dr. Krestin, who is also on FLUIDDA’s board of directors. “As a result, one could theoretically show the same image to three radiologists and sometimes receive a different interpretation from each.”  

 (FLUIDDA is the world leader in the field of Functional Respiratory Imaging, which marries HRCT scans and Computational Fluid Dynamics technology offering vast improvements by making clinical trials shorter, faster and more cost effective. Its advanced engineering tools are applied to improve the diagnosis and monitoring of respiratory diseases.) 

 At ZNA Cadix an internal investigation launched in the fall of 2023 revealed that “the quality of medical imaging did not always meet the strict quality standards.” 

 Concerns about the performance of the radiologists were brought to Dr. Bervoets’ attention from doctors both inside and outside of the hospital. 

 According to the Brussels Times, the investigation reviewed 1,300 cases identifying 52 “as potentially influencing patient safety, prompting the hospital to contact 24 patients.” 

 VRT, the national public broadcaster for the Flemish Community of Belgium, reported that errors included “a bone fracture that was missed, a dilation of a blood vessel that was not noticed, or the size of a tumor that was measured incorrectly.” 

 While ZNA Cadix scrambled to contain the fallout, it made one point clear, that the mistakes were made by humans, not machines.  

 “The quality of the devices in ZNA Cadix is beyond dispute,” ZNA said. “These are the newest devices on the market.”