Lung cancer care in the COVID-19 aftermath. How AI can support radiology teams
By Catalina Barzescu
Healthcare practitioners continue to deserve our applause for their immense dedication taking care of COVID-19 patients over the last months. At the same time, as many countries are establishing a new ‘normal’, in which the virus is present but fairly under control, we have the headspace to reflect on the impact of the pandemic and prepare for future challenges.
Early June, we attended the British Institute of Radiology (BIR) virtual event on COVID-19 imaging. Aidence’s Lizzie Barclay and Jeroen van Duffelen joined a great lineup of speakers who shared lessons learnt from dealing with the disease and perspectives on post-pandemic developments. In the context of amplified challenges, we showed how AI can support radiologists dealing with the post-COVID lung cancer workload. (You can view the complete talk here.)
COVID-19 and the lung cancer pathways
COVID has brought dramatic changes in all healthcare pathways. A general response from medical institutions has been increasing the capacity for COVID and emergency patients, by reducing capacity for routine healthcare practice. Patients themselves have also been avoiding seeking medical advice or presented late. As a result, the impact on non-communicable diseases, such as cardiovascular disease, diabetes, and cancer, appears to be significant; time and research will show just how significant.
In lung cancer care, COVID-19 has led to a reduction in all pathways leading to diagnosis. Surgeries and treatments have been postponed. Lung cancer screening, in particular, has been paused or stopped completely; thousands of early lung cancers are thus ‘missed’, and patients are expected to present at a later, probably less treatable stage. Patients’ fears of going to a hospital due to the perceived presence of infection, as well as the economic impact of unemployment or insurance loss, may be further obstacles to cancer screening. In the UK, as an example, an estimated two million people are waiting for cancer screening and treatment.
Overall, the impact will most likely be poorer survival prognosis for patients at risk or diagnosed with the disease. At the BIR event, professor Muntzer Mughal, Clinical Lead for the North Central and East London Cancer Alliance, showed a model of additional demand for lung cancer-related activity going well into 2021. Lung cancer care post-COVID may just be a ‘crisis after the crisis’.
From normalization to acceleration
Besides the patient impact, we must consider the strain on the medical staff. COVID cases are still being reported, yet not at peak levels, a trend that will hopefully be maintained. Radiology teams are gradually returning to ‘normal’ practice — one with the added precautions of equipment cleaning and social distance to minimize further infections.
In the medium-term, radiology departments will have to go through the non-urgent scans that have been postponed or rescheduled due to COVID, a volume of work that will create new time pressures or stress for radiologists. In some countries, normalisation efforts will further face a shortfall of modalities (such as the UK’s relatively low number of scanners per capita).
On the longer term, we foresee an ‘acceleration’ phase. Once normalisation has resumed, healthcare systems will want to make changes that had been in place before COVID. Lung cancer screening, for instance, will likely ramp up. The NHSE Long Term Plan for lung health checks set a target to diagnose 75% (vs. the current 50%) of all cancers at an early stage (1 or 2) by 2028, so there might be a push to meet these targets. Throughout these developments, healthcare systems will also maintain a reserve capacity to cope with possible next COVID waves.
Healthcare practitioners who have been under pressure since the onset of the pandemic may, therefore, be asked to deal with an increased workload post-COVID, testing their resilience and own health.
AI in the new ‘normal’
The above considerations, as well as the many gaps COVID has highlighted within healthcare systems, accounts for extra demand to bring in technology that can improve healthcare systems. Artificial Intelligence (AI) presents opportunities to augment health by taking on some of the time-consuming tasks from human hands, and contribute to economic benefits and better patient care. Initiatives such as the NHSX AI lab recognise this potential.
Following numerous discussions we held with healthcare institutions and radiologists over the past months, we see several roles for AI in the post-COVID lung cancer pathway:
- AI systems are well-equipped to spot subtle patterns on medical images. Having a second pair of eyes when analysing chest CTs can support incidental nodule detection, thus picking up early lung cancers. This increased capacity may provide the balance we are lacking because of screening programmes being postponed.
- Nodule volumetric measurements with AI can speed up decision making by automatically providing insight into nodule growth rates.
- AI-driven productivity increase would allow radiology teams to stay on top of reporting of the non-urgent scans, within the turnaround time targets.
- AI features can be developed and tailored to maximise efficiency in processing screening scans. For example, we added an interactive reporting module to Veye lung Nodules, to support radiologists with the detailed reporting protocol. For a more in-depth look at AI for lung cancer screening, read Lizzie Barclay’s recent article.
Within all use cases, AI can and should facilitate remote working patterns. One of the takeaways from the COVID crisis is that remote radiology allows for effective reporting while maintaining social distance; allowing radiologists the option to work from home post-COVID would increase flexibility in their workday. For AI solutions not to block remote work, they must fully integrate with the radiology workflow and system. Now more than ever, AI should be agnostic to PACS.
The reset button
The British Institute of Radiology referred to COVID-19 as the ‘reset button’ for radiology. The question is: how can radiology emerge even stronger post-COVID? Now is a time to seize in preparation for the expected post-COVID scan volumes and requirements. Leveraging the power of technology can be one of the opportunities to do things differently in what will be a different order of things.