Applications of artificial intelligence (AI) in kidney cancer, although still in the early stages, have shown promise in a variety of clinical applications, noted the authors of a review on the topic in ASCO Educational Book.
“The diagnosis, characterization, treatment, and management of kidney cancer is ripe with areas for improvement that can fulfill the promise of artificial intelligence,” wrote Ivan Pedrosa, MD, PhD, of the University of Texas Southwestern Medical Center in Dallas. and co-authors. “Here we examine the impact of current AI research work for clinicians caring for patients with kidney cancer, focusing on the perspectives of radiologists, pathologists and urologists.”
In the following interview, Pedrosa, professor of radiology and vice chair of radiology research, discussed some of the promising applications as well as the challenges that need to be overcome.
What is one of the most pressing unmet clinical needs in kidney cancer that AI could help address?
Pedrosa: I would emphasize the treatment of early stage kidney cancer. Most kidney cancers today are diagnosed when a small kidney mass is found incidentally during an imaging study that was performed for an unrelated medical condition, such as abdominal pain, trauma, etc. These masses are often benign – one in five – and when malignant, can represent one of many different types of cancer. Currently, treating these small kidney masses is difficult because we have no reliable way to tell what they represent or how they will behave, even with a tissue biopsy.
Ideally, we would like to avoid unnecessary treatment of benign and indolent malignancies—the latter especially in patients with competing comorbidities and limited life expectancy for whom active surveillance may be an option. However, we also want to speed up the treatment of aggressive malignant tumors that can threaten the patient’s life.
Early evidence suggests that artificial intelligence has the potential to play a role in helping both doctors and patients make these decisions. Specifically, AI can improve the characterization of these masses in imaging studies and offer prognostic information that is not well captured in renal biopsy analyzes today.
In what areas of kidney cancer diagnosis and treatment have AI algorithms performed well?
Pedrosa: Artificial intelligence is in its infancy when it comes to applications in kidney cancer. Some promising results have been reported in the prediction of histological subtypes based on the analysis of imaging studies such as computed tomography and magnetic resonance imaging with deep learning algorithms.
There is also exciting data using AI algorithms to predict molecular changes, such as mutations, using digitized histology images. The performance of some of these algorithms is remarkable and, if validated, could have a substantial impact in clinical trials and ultimately in clinical practice.
In what areas did they also fail?
Pedrosa: Artificial intelligence algorithms perform well in the repetitive tasks they are trained for. Unfortunately, it is difficult to train algorithms to deal with the huge variability we see in the clinic. For example, while AI algorithms perform extremely well for automatic segmentation of the kidney because the shape of the kidney is more or less predictable, their performance drops for segmentation of real kidney tumors, which are more variable in size, shape, and appearance. However, we will definitely continue to see progress in this area.
What obstacles need to be overcome for artificial intelligence to reach its full potential?
Pedrosa: Artificial intelligence algorithms need adequately edited and annotated data for training. Their performance suffers when the training data is inconsistent or when the algorithm encounters scenarios that were not included in the training dataset. Unfortunately, there are countless sources of variability in kidney cancer patients, and training AI algorithms for them would require huge datasets. Examples include differences in patient body habitus, imaging technique, image acquisition protocols, histological subtype, tissue processing, molecular clusters, etc.
The use of AI algorithms in clinical practice is likely to expand when we figure out a way to generate large, well-edited data sets that are diverse enough so that AI algorithms can be trained to recognize most clinical scenarios. Meanwhile, AI algorithms can provide workflow improvements to make some of the repetitive and time-consuming tasks that humans do, such as image annotation, easier.
What do you think will be the role of AI in the diagnosis and treatment of kidney cancer in 10 or 20 years?
Pedrosa: I believe that AI will be at the center of deciphering tumor heterogeneity – probably the most difficult challenge in kidney cancer today, and perhaps oncology in general. In recent years, a number of new and exciting therapeutic regimens have been approved for patients with locally advanced and metastatic disease. Fortunately, we’ll continue to see more options become available.
At present, however, the choice between these regimes is largely empirical. Although new data suggest that certain molecular signatures may indicate a higher likelihood of response to some of these regimens, this determination requires tissue samples.
Unfortunately, we know that kidney cancer is highly heterogeneous and it is quite possible that a single biopsy is not representative of the entire tumor burden, at least in some patients. Although we cannot obtain tissue from every metastatic site, other approaches such as imaging or liquid biopsies have the potential to evaluate the entire tumor burden. This is extremely possible even today when we combine systemic therapy with local interventions such as surgery, ablation or radiation in patients with a heterogeneous response.
However, the amount of information we need to process to make these decisions and evaluate oncology outcomes at the patient and population level is staggering. I believe that artificial intelligence will play a key role in integrating all these data – imaging, pathology, laboratory, molecular, omics, etc. – to develop new multidimensional predictive and prognostic biomarkers that will help in the treatment of kidney cancer patients from their early diagnosis of metastatic disease.
Read the study here and expert commentary on the clinical implications here.
#Ivan #Pedrosa #PhD #applications #artificial #intelligence #kidney #cancer