Patient and Treatment Monitoring

In cancer care, monitoring helps check if treatments are working and looks for signs that the disease is getting worse. This can be done through tests like biopsies, scans, or noticing a patient’s symptoms getting worse. These tools for monitoring are widely used, but they are not always accurate or consistent. Our research aims to fix these problems by using AI technology to make monitoring more precise and reliable, which can lead to better care and improved outcomes for patients.

Is the treatment working?

How can we decide if a treatment is working? One option is to use a response evaluation criteria. Response evaluation criteria are a set of rules or guidelines that doctors use to measure how well a cancer treatment is working. These criteria help determine if the tumor is shrinking, staying the same, or growing. They are like a “report card” for cancer treatment, showing whether the treatment is effective or if adjustments are needed. For example: If a tumor gets smaller, it might be called a partial response; if it disappears, it’s a complete response; and if it grows, it could be classified as a progressive disease. Doctors use these evaluations during treatment to decide the next steps, such as continuing, changing, or stopping a treatment.

What is wrong with the current response criteria?

This research stems from the current doctoral works of Teresa Tareco Bucho.

Current methods for measuring treatment response, like RECIST (Response Evaluation Criteria in Solid Tumor), lack accuracy and consistency. Teresa’s work explored how factors that define the RECIST criteria (selection of target and measurable lesions, objectivity/subjectivity in the selection of target lesions, and timing of the evaluation) affect the variability of the assessments. She further studied the feasibility of alternative methods, including blood markers and total tumor volume. Our findings aim to guide the development of more accurate and reproducible response criteria in the future.

Automated Response To Treatment In Mesothelioma (ARTIMES)

This research stems from the MSc and PhD thesis of Kevin Groot Lipman, the current doctoral work of Valerio Pugliese, and in collaboration with the Department of Thoracic Oncology.

Mesothelioma is challenging to measure because, unlike round tumors, it spreads along the lung’s surface, creating crescent-shaped growths. Traditional methods rely on measuring the size of tumors with straight lines (diameters), which doesn’t work well for mesothelioma’s irregular shape. The ARTIMES criteria use AI to analyze CT scans and label every part of the image as either tumor or non-tumor tissue. This allows it to calculate the total tumor volume, which gives a much clearer picture of how the disease is progressing or responding to treatment. Previously, measuring total tumor volume manually was too time-consuming to be practical. ARTIMES makes this process faster and more accurate, even identifying when the cancer worsens weeks earlier than current methods. This means doctors can adjust treatments sooner, avoiding unnecessary treatment related side effects and potentially improving patient outcomes. By automating tumor measurement, ARTIMES is transforming how mesothelioma is monitored and treated, making it easier to track this hard-to-measure cancer.

Automated Response To Treatment In Neuroendocrine Tumors (ARTINET)

This research stems from the current doctoral works of Kalina Chupetlovska, the current master thesis of Jacco Engel, and in collaboration with the Department of Medical Oncology.

Building on the ARTIMES project, we launched ARTINET: tracking treatment responses in gastroenteropancreatic neuroendocrine tumors (GEP-NETs). GEP-NETs are a diverse group of tumors originating from neuroendocrine cells within the gastrointestinal tract and pancreas. Traditional methods for assessing treatment response in these tumors often face challenges due to their varied presentation and behavior. Kalina and Jacco’s work aims to develop an AI-driven approach to enhance the precision and consistency of response evaluations in GEP-NET patients. By analyzing total tumor volume, the new criteria aim to provide more accurate assessments of how tumors respond to therapies, potentially leading to improved patient management and outcomes.

Automated Response To Treatment In Soft Tissue Sarcoma (ARTISARC)

This research stems from the current doctoral works of Iris van der Loo.

Unlike mesothelioma and neuroendocrine tumors, where tumor growth indicates treatment inefficacy, some tumors display different behaviors. Soft tissue sarcomas are one of these tumors. Iris investigates the application of AI to enhance the evaluation of treatment responses in soft tissue sarcoma. Sarcomas encompasses a diverse group of cancers originating from connective tissues, and assessing their response to therapy can be challenging due to their heterogeneous nature.

Bladder Cancer Response Evaluation (BLARE)

This research stems from the current doctoral works of Joyce Greidanus, and in collaboration with the Departments of Medical Oncology and Urology.

Primary bladder tumors can often develop with irregular shapes, just like mesothelioma, but on a smaller scale. Joyce’s work focuses on developing a set of criteria for primary bladder cancer, with particular attention to the differences in assessments made by various radiologists, and the use of AI in mitigating variability that can occur when interpreting MRI results. In current clinical practice, the bladder is typically removed after systemic therapy. With promising results from new systemic treatments that lead to complete response in some patients, organ-sparing approaches can be explored. With the use of AI, we are developing new criteria to assess treatment response in bladder cancer.

Image-to-image registration and voxel-wise tracking

This research stems from the current doctoral works of Laura Cerquin Estacio.

Tracking total tumor volume marks a significant advancement over the current clinical standard. However, for cases involving complex tumor dynamics, a single measurement may not be sufficient. Laura’s work focuses on developing more refined methods for monitoring tumors and patients during treatment. Her approach aims to go beyond simply measuring total tumor volume, striving to analyze and track changes at the level of every single voxel.

Avatar of Kalina Chupetlovska Avatar of Laura Estacio Cerquin Avatar of Joyce Greidanus Avatar of Kevin Groot Lipman Avatar of Valerio Pugliese Avatar of Teresa Tareco Bucho Avatar of Stefano Trebeschi Avatar of Melda Yeghaian Avatar of Iris van der Loo

Publications

show more