Looking to the Future of Early Detection in Cancer: Liquid Biopsies, Imaging, and Artificial Intelligence

Abstract

Cancer is a highly heterogeneous disease that is frequently diagnosed at an advanced stage with the presence of metastatic disease resulting in limited treatment options and a devastating prognosis. The early detection of cancer is essential to achieving better patient outcomes and reducing mortality rates. Liquid biopsy, also known as a fluid phase biopsy, consists of the assessment of liquid biological specimens, most commonly blood. Liquid biopsy represents a minimally invasive alternative to conventional tissue biopsies with the potential to significantly improve cancer diagnostics and management. It is a highly sensitive approach which offers the ability to detect circulating DNA (cell-free tumor DNA [ctDNA]/cell-free DNA including ctDNA [cfDNA]), exosomes, and/or proteins to provide a comprehensive snapshot of the heterogeneous tumor status and environment. Furthermore, liquid biopsy can identify genetic alterations and disease-related biomarkers, giving clinicians valuable insights into tumor biology, the staging of cancer, and the ability to monitor and adjust treatment regimens. This wealth of information supports the development of personalized therapeutic strategies, thereby improving patient outcomes. However, liquid biopsy alone is not enough to provide crucial information about cancer development. Two important clinical modalities that can complement liquid biopsy assessment are imaging and the application of artificial intelligence (AI). Imaging technologies such as magnetic resonance imaging (MRI), computed tomography (CT), and positron emission tomography (PET) have long been used for diagnosis and monitoring of patients with cancer. These imaging techniques have dramatically evolved to offer higher spatial resolutions, better contrast, and improved functional assessments of tumors. These advancements aid in the early detection and localization of malignancy, determining tumor size and aggressiveness, and the identification of metastatic spread. In contrast, AI involves using machine learning models to analyze medical data from multiple sources to yield additional insights for improved patient care management. Using AI-driven algorithms alongside imaging technologies may optimize the interpretation of complex image data and minimize inter-observer variability. Combining liquid biopsies, imaging, and AI applications can significantly enhance cancer diagnostics and management. This would enable early detection of cancers with liquid-phase testing, the production of a detailed spatial and functional assessment of disease using imaging, and determining risk management and healthcare outcomes with the pairing of AI. Despite the complementary nature of these techniques, their combined application is still at an emerging stage in clinical oncology. Nevertheless, these modalities have great promise and could become potential game changers for the early detection of cancers. In this Q&A article, internationally recognized clinical and scientific experts discuss the growing implications, benefits, challenges, and future potential of liquid biopsies, imaging, and AI in oncology to enhance patient management and care. Kenneth Maiese: The challenges for diagnosing and treating solid tumors are multifold. Current methods for cancer detection rely upon scheduled examinations or the onset of clinical symptomatology that would involve physical examination, cellular analysis such as a pap smear, imaging such as computed tomography, mammogram, and procedures that involve biopsy of tissue and possible direct visualization such as with colonoscopy. These methods can have limitations, especially since the onset and development of tumorigenesis as well as disease recurrence may occur much earlier in the body beyond the detection limits of these methods that would allow for tumor progression and potential metastatic disease. The use of tissue biopsy for diagnosis also is not without imperfections since tissue biopsy can be invasive, costly, and yield samples that may not be entirely conclusive due to heterogeneity of the tumor sample, sample contamination, and the presence of apoptotic cells with additional necrotic and fibrotic tissue. In regard to cancer treatment, this can be compromised by antigen escape, inability to infiltrate the tumor mass with chemotherapeutic agents due to poor vascularization, radiotherapy that may lead to immunosuppression, and tumor toxicity with cytokine release. Christian Rebhan: Diagnosing and treating solid tumors is not always straightforward. At Unilabs, we use state-of-the-art technologies including AI. Even so, in their early stages, solid tumors are not easy to detect due to lack of clinical signs and symptoms. At diagnosis, these tumors may have grown larger and metastasized, complicating disease management. Early involvement of lab diagnostic solutions is desirable to facilitate even earlier detection. Another challenge is that solid tumors can exhibit significant heterogeneity (populations of genetically diverse subclones). Today, this requires optimal focus on the biopsy process to allow for efficient sampling of representative tissue, followed by tissue processing and classical histopathological analysis. Where molecular profiling is needed, the analysis is carried further through cell/tissue extraction molecular testing. This blend of traditional, time-tested histological methods with recent addition of molecular analysis like next generation sequencing is complicated, but works well in laboratories with sufficient experience and expertise. Monitoring patients while undergoing cancer therapy as well as during remission poses a third challenge. Patients experience different responses to cancer therapy and chemotherapy, and radiation therapy, and even immunotherapies can cause significant side effects. It is important to maximize the desired response to therapy while minimizing side effects via constant monitoring and adapting the therapeutic approach accurately to match the disease status. Monitoring solid tumors needs a delicate balance between repeated sampling and clinical judgement to enable disease management. For patients in remission, the question is how to best balance follow-up visits and investigations for asymptomatic patients vs waiting and hoping for no signs of recurrence. Subba Digumarthy: There are several challenges with the current diagnostics and treatment options for solid tumors. The outcome is closely tied to the cancer stage: an early-stage diagnosis brings the best prognosis. Few solid cancers such as breast, cervical, colon, and lung cancer have approved and effective screening tests, with variable adoption across the vulnerable or at-risk population. In these screening tests, such as in the case of lung cancer screening with low-dose CT and colon cancer with CT colonography, often the high sensitivity comes at the cost of low specificity and risk of detecting incidental findings. Although therapeutic options for solid cancers have improved, multiple time point imaging and genomic testing is often needed to determine the standard treatment and monitor treatment response. In advanced solid cancers, the outcome remains poor despite the increase in treatment options. Joan Anton Puig-Butille: There are various clinical scenarios where liquid biopsy tests can significantly enhance patient care and guide clinical practice. Tests based on ctDNA detection are particularly valuable in patients where the tissue sample is unavailable or inadequate in quantity and quality for conducting molecular studies. In such circumstances, liquid biopsy enables the acquisition of the tumor's molecular profile and facilitates the identification of patients who may benefit from precision oncology approaches such as targeted therapy treatments. These tests are also relevant within the context of patient monitoring. As a minimally invasive test, it allows for continuous surveillance of the cancer patient, enabling the monitoring of disease evolution and treatment response. The levels of ctDNA do correlate with the response to any type of systemic treatment. In addition, these tests might allow the identification of underlying resistance mechanisms that explain the lack of treatment response. Other promising aspects of liquid biopsy in cancer applications are the detection of minimal residual disease in patients undergoing surgical resection of the primary tumor, early diagnosis, or cancer screening. However, the use of liquid biopsy tests in these contexts still poses significant challenges that need to be addressed. Subba Digumarthy: Minimally invasive liquid biopsy tests can help improve the specificity of other early screening tests such as low-dose CT for lung cancer screening or mammography. Often these tests can be integrated with imaging-based screening tests to determine the risk and probability of indeterminate findings such as lung nodules and polyps in colon and mitigate patient anxiety and avoid unnecessary follow-up imaging for benign lesions. For patients with suspicious lesions, this can expediate the workup for early diagnosis. Kenneth Maiese: Cancer continues as the leading cause of death worldwide with at least 2 million cases globally. Interestingly, several factors have contributed to reduced cancer mortality over the prior decade that include greater patient awareness of cancer screening programs, improved imaging, surgical, and radiotherapy procedures, and new treatment strategies that encompass chemotherapy, immunotherapy, and precision medicine. Yet, an important void in cancer care continues to exist that liquid biopsy can offer by providing the early, specific, and sensitive detection of tumor onset or recurrence that will target early primary tumor treatment and limit or abolish subsequent metastatic disease. As current research continues to progress, several present observations highlight the potential utility of liquid biopsies. First, when tumor cells become present in the blood, liquid biopsies can be the first line in the detection of a developing cancer that allows for the implementation of early and aggressive treatment protocols to overcome tumor growth. Second, liquid biopsies may be able to uncover cancer lymph node invasion and distant metastasis prior to detection with imaging or biopsy, providing the capability to monitor cancer progression and the outcomes of patients more closely. Third, liquid biopsies can identify early mutational changes in cancers that otherwise would be undetectable in a corresponding biopsy, allow initial detection of resistance to treatment, and provide a sensitive process to monitor treatment efficacy. Christian Rebhan: Liquid biopsy tests can solve some important challenges associated with solid tumor diagnosis and clinical management. Their use depends on the scientific characteristics of the cancer type, cancer stage, and therapy availability. Furthermore, practical factors like test viability, access, and reimbursement should also align to enable scaling of such tests. Examples for use include, but are not limited to, early detection and noninvasive screening. Liquid biopsy tests can identify ctDNA or tumor-genetic mutations before clinical symptoms appear, thereby promoting early detection and improving patient outcomes. Such tests can also be used for screening of suitably targeted populations. Another area of use is the monitoring of treatment response. Liquid biopsy tests can be employed to monitor treatment effectiveness in cancer management. By analyzing ctDNA, the tests provide real-time information on tumor dynamics and treatment response. This allows healthcare providers to assess if a treatment is working or if adjustments are necessary, such as switching to alternative therapies or adjusting drug dosages. Liquid biopsies also allow for genetic profiling of the tumor composition including mutations that generate spontaneously or are acquired due to therapy. This information is useful for identifying disease prognosis, patient stratification, outcome prediction, and guiding treatment decisions using targeted therapies. Post-treatment patient surveillance and detection of minimal residual disease (MRD) are further use cases coming to mind. Joan Anton Puig-Butille: The field of liquid biopsy is evolving rapidly. However, there are still technical issues that limit its clinical applicability, especially in the context of early detection. Therefore, liquid biopsy and tissue biopsy should be considered complementary strategies rather than exclusive ones. To use liquid biopsy strategies instead of tissue biopsy, we need evidence in terms of analytical validation, clinical validation, and clinical utility to establish the superiority or equivalence of liquid biopsy compared to tissue biopsy in specific clinical settings. Interestingly, recent approaches based on epigenetics or fragmentomics of cfDNA show promising results and might improve the sensitivity and specificity of ctDNA-based tests in certain clinical contexts. Finally, it is important to conduct comparative studies and randomized trials demonstrating the clinical benefits of these liquid biopsy approaches prior to their implementation in the clinical setting. Christian Rebhan: One key advantage of liquid biopsy tests is convenience and patient experience due to their noninvasive nature vs solid tissue biopsy. Their use should be context-specific for the tumor type and considered an alternative form of testing that is accessible to the patient. Criteria I considered of critical importance for the use and uptake of liquid biopsy tests include their analytical validity (sensitivity, specificity, reproducibility, and robustness of the liquid biopsy assay), their clinical validity (positive predictive value, negative predictive value, and comparative concordance data), and subsequently clinical utility (impact on patient outcomes, treatment decisions, survival data and quality of life indicators). Once these criteria have been sufficiently established, expert consensus and endorsement by professional bodies, as well as the economic viability and ease of technical implementation in communities, will be important for broader adoption. Subba Digumarthy: Most liquid tests suffer from low sensitivity despite their high specificity. Because there is only a miniscule amount of circulating tumor protein or tumor-related nucleic acids, they require a larger volume of blood for testing. Currently the accuracy of liquid biopsy is higher with a large tumor burden. The purpose of biopsy is to confirm the diagnosis, establish histological subtype, and determine the underlying genetic alteration to choose the treatment. So, in most cases, tissue biopsy should be performed, if safe. There are instances where liquid biopsy can be a suitable substitute such as: (a) a lung nodule with suspicious imaging features that is risky to biopsy, (b) recurrent cancer after treatment, where liquid biopsy can be a suitable substitute for tissue biopsy, and (c) in patients enrolled in clinical trials who need repeat biopsies while on treatment. Kenneth Maiese: I believe that there exist different scenarios for the use of liquid biopsies for the detection of cancer, but do not foresee at this time liquid biopsy as a replacement for tissue biopsy. Liquid biopsy, a process derived from the direct contact with the tumor microenvironment, offers the ability to frequently, noninvasively, and cost-effectively detect tumor onset, monitor treatment response, and assess prognosis during disease progression. Liquid biopsy has a significant advantage when tumor detection is beyond the limits of current physical examination, direct visualization, or radiographic imaging techniques. The caveat is for liquid biopsies to have appropriate sensitivity and specificity, especially with tumors that have complex cellular pathophysiology and reduced levels of detectable circulating protein or genetic material. In other scenarios with tumor detection on imaging and attainable tissue samples, liquid biopsy testing can be a welcome counterpart for the diagnosis, prognosis, treatment, and surveillance of tumors with additional necessary information, since tissue specimens may be incomplete as a result of tumor heterogeneity and sampling limitations. Christian Rebhan: Whereas liquid biopsy provides molecular-level genetic information, imaging provides anatomical and functional data to localize the tumor and understand behavior. We have many countries in our network that utilize such an integrated approach using lab results, imaging, pathology, and genetic data for managing specific cancer types such as breast and prostate. Imaging is mandatory for staging, understanding lymph node/organ involvement, and metastasis. Liquid biopsy can identify genetic mutations in the tumor, both at primary and secondary sites. Integrating the information from these modalities can improve diagnosis as well as treatment planning, including use of targeted therapies. Both techniques can be used in a complimentary fashion to adapt treatment to disease progression. Periodic imaging scans post treatment can be combined with liquid biopsy results to guide patient management with regard to recurrence, relapse, and MRD evaluation. Joan Anton Puig-Butille: Imaging plays an important role in association with liquid biopsy in clinical care. While liquid biopsy provides molecular and genetic information about the tumor, imaging techniques such as CT, MRI, and PET complement this data by providing visual, molecular, metabolic, and anatomical information about the tumor and its surrounding tissues. The information obtained by imaging techniques is crucial for accurate staging, treatment planning, and monitoring of disease progression. The integration of imaging with liquid biopsy in clinical care might provide a more comprehensive and multidimensional understanding of cancer, enabling more precise diagnosis, treatment selection, and monitoring of patients. Kenneth Maiese: Both liquid biopsies and imaging independently bring vital attributes for the diagnosis and treatment of solid tumors. Liquid biopsies can offer significant benefits as a noninvasive tool; they require no radiation exposure to the patient, they offer early detection that goes beyond traditional means with clinical examination and imaging, and they can be obtained frequently, rapidly, in real-time, and are cost-effective. Yet, when tumor presence is detectable through imaging, imaging provides localization of the primary tumor site as well as the presence and location of metastatic tumor regions. Visual inspection through imaging is necessary for the staging of solid tumors and that is not available using liquid biopsy testing alone. Subba Digumarthy: Imaging and liquid biopsy can complement each other in clinical care. While imaging studies have high sensitivity for detecting macroscopic solid cancer, liquid biopsy tests offer high specificity. Thus, a combination of imaging and liquid biopsy can help address their limitations and create a highly sensitive and specific diagnostic pathway for oncologic care, which does not involve invasive tissue biopsy. Subba Digumarthy: AI can help liquid biopsy expand its applications and improve diagnostic performance by enabling rapid analysis of several hundred proteomic signatures of cancer-related circulating proteins and nucleic acid molecules. Likewise, imaging studies which are now interpretation can become or such as with AI can imaging interpretation to the next or with or without and other which is not possible in the current assessment of solid cancer. AI models can bring proteomic and genomic information from liquid biopsy and imaging information to a more diagnostic pathway that could help benign and lesions, unnecessary additional or follow-up imaging, and the need for invasive tissue biopsy. Joan Anton Puig-Butille: analysis strategies a key role in the analysis of integrated liquid biopsy and imaging we have the technical to a large amount of relevant data from a liquid biopsy, including molecular information from cfDNA or ctDNA, as well as information from other such as circulating tumor cells or advanced imaging techniques, such as allow for the extraction and analysis of features from medical As we increase the amount of complex we require artificial intelligence strategies to that the of information and imaging For image data can be combined with liquid biopsy data and using AI algorithms to predictive models for treatment response, prognosis, and therapeutic Christian Rebhan: AI and machine learning can the data by imaging and liquid biopsies, within each as well as in approaches of integrated AI algorithms analyze CT, MRI, and PET scans to in tumor and This of tumor identifying and certain Combining this data with liquid molecular can help more comprehensive of the even a AI can facilitate diverse data from imaging, liquid biopsy, and genetic tests, alongside clinical information from the This can help generate comprehensive patient specific to each patient, and help personalized This approach to clinical diagnostics information for clinicians managing and for their patients. Kenneth Maiese: AI using machine learning models can offer a significant benefit to the data obtained from liquid biopsies and data from liquid biopsies and imaging, especially in the of proteomic and genetic data at the cellular during tumor present a challenge to identify predictive with the use of only standard AI can analyze complex of data that are and to generate predictive signatures for clinical care. Although further studies multiple patient are necessary for the of machine learning of information from liquid biopsies and imaging could the rapid interpretation of and imaging information that can precision and offer providers clinical predictive models for effective care treatment ctDNA, circulating tumor CT, computed The corresponding that on this have the criteria of for (a) significant to the and acquisition of or analysis and interpretation of (b) or the for (c) of the and to be for aspects of the that to the accuracy or of any of the are and who for has been from the the as an of and from of from and as an of is an of provides image analysis for clinical research trials for and has research from and and for from and of the of a diagnostic imaging, and We would like to the of of and and in the of this

Topics & Concepts