Biomarkers in Cancer Screening: Promises and Challenges in Cancer Early Detection

Introduction:

Cancer is a major public health and economic burden worldwide with an estimated 19.3 million new cancer cases and 10 million deaths in 2020, approximately accounting for one in six deaths [1]. In 2023, a total of 1.9 million new cancer cases and 610,000 cancer related deaths are projected in the USA, indicating slight decrease in the overall cancer mortality [2]. Detecting and treating cancer at early stages enables more effective treatment and reduces the morbidity and mortality [3]. 5-year survival rate across all cancers reduces from 89% when localized to 21% once metastasized [4]. For example, in colorectal cancer, the 5-year survival for stage I is 80–95%, but it is less than 15% when detected at stage IV. Likewise, the 5-year survival for resectable stage IA pancreatic cancer is about 85%, but less than 10% when detected at an advanced stage [5,6].

Although there is inconclusive evidence that PSA-based screening results in a decrease in prostate cancer (PCa) mortality [7–10], early detection has resulted in significant increase in the number of men diagnosed with localized and locally advanced disease. Since 2014, PCa incidence has increased by 3%/year, mostly driven by 4–5%/year increase in incidence of regional and distant stage diagnosis [11]. Randomized clinical trials (RCTs) investigating PSA-screening have demonstrated a reduction in the incidence of de novo metastatic PCa, but this has not been studied in large, population-based studies [12].

Additionally, costs associated with treating cancers diagnosed at a late-stage are up to 7 times higher than those diagnosed at an early-stage, leading to significant economic burden [2]. Early detection includes early diagnosis and screening. According to the US Preventive Task Force (USPSTF), single-site cancer screening is recommended for colorectal, breast, cervical and lung cancer for at-risk individuals (Fig. 1) [13–17]. Early cancer detection through established screening/diagnostic approaches has now led to cervical, breast, and colorectal cancers being diagnosed sooner, at earlier stages, than cancers without established screening modalities, which constitute almost 70% of the cancer deaths in US [2]. Many cancers such as pancreatic, esophageal, and ovarian cancers are often diagnosed at advanced stages leading to significant morbidity and mortality.

Figure 1:

USPSTF- Recommended Screening Tests for Cancer

USPSTF- United States Preventive Services Taskforce; created by Biorender.com

Thus, there is a critical need to identify cancer specific biomarkers which in conjunction with imaging methods can improve screening/diagnosis of cancers with no established standard of care method. The need for continued research also persists for those cancers where the currently available biomarkers do not have optimal clinical performance like the PSA test for prostate cancer. According to the National Cancer Institute (NCI), a biomarker is “a biological molecule found in blood, other body fluids, or tissues that is a sign of a normal or abnormal process, or of a condition or disease [18] and are objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention [19]. The rationale for selecting a biomarker in the cancer field includes its ability for predicting risk, early detection, staging, monitoring treatment response and tracking disease progression. This is in addition to a biomarkers ability to reduce over-diagnosis, distinguish aggressive from non-aggressive lesion types, and inform selection of patients for specific treatment options, thereby leading to greater life expectancy and quality of life for the patient [20]. The characteristics of an ideal biomarker are high specificity, sensitivity, quantifiability, ease-of-use, reproducibility, clear and concise read-outs for clinicians, cost-effectiveness, and can be measured from easily acquired biological fluid or specimen. The present review primarily addresses the current state of knowledge of available early detection molecular markers in cancer, new developments and technologies in the field and discusses their challenges and potential for clinical utility. Imaging modalities are not discussed here.

Current Screening and Early Detection Modalities:

We have used the phrases ‘early detection’ and screening’ interchangeability. However, screening is a process involving multiple modalities, e.g., biomarkers, imaging, etc. that looks for cancer before symptoms appear. Early detection, on the other hand, looks for the earliest stage of disease that is clinically manageable. Screening is improved for cancers with an established early asymptomatic phase and available clinically validated, safe, sensitive, specific, and straightforward screening test with strong patient adherence, acceptance by patients and clinicians, and cost effectiveness.

United States Preventive Services Taskforce (USPSTF) recommended screening tests for breast, colon, lung, prostate, and cervical cancers along with their clinical performance is shown in Fig. 1. Screening is more effective for slow growing tumors than rapidly growing tumors due to lead-time bias. Yet, screening suffers from overdiagnosis of certain cancers, leading to overtreatment or unnecessary diagnostic biopsy procedures. Further, in relation to the clinical performance, most of the screening tests have sensitivity and specificity in ranges of 70%−80% and 60–70% respectively [21], low PPV [22,23] and high overall false positive rate when the test is used sequentially [24] and therefore is not a perfect solution to early detection. Additionally, there is bias effect of disease prevalence on PPV; for cervical cancer with disease prevalence of <0.1 %, PPV is <0.1%, while for high-risk lung cancer with disease prevalence of 1.1%, PPV is 6.9%. Other limitations are compliance with recommended screening tests for example compliance is in range of 69–80% for breast, cervical and colorectal cancers while it is 5% for lung cancer [25–29] and lower cancer screening rate within underrepresented racial/ ethnic populations [2,30]. Additionally, biological and technical challenges of endogenous markers including tumor heterogeneity, interpatient variation, comorbidities and background signal by healthy cells, remain obstacles for the early detection of cancer specially before the symptoms appear [31]. Therefore, we must explore new multi-analyte tests which incorporate new quantitative imaging modalities with adjunct biomarker/ screening assays, which can be broadly applied across patients with different ancestries.

Types of Early Lesions

In response to the Cancer Moonshot Initiative’s Blue-Ribbon Panel recommendation, NCI launched Human Tumor Atlas Network (HTAN) program with the overall goal of generation of human tumor atlases. The results of this initiative are helping towards defining the precancer framework, which should be dynamic and adaptable with following important interrelated components: histopathological features, biology in conjunction with disease outcome, changes in the precancer microenvironment (PME), and molecular features derived from the integration of the multi-omics data [32]. Precancerous lesions are high-risk lesions/states which often lead to the development of invasive carcinoma (Fig. 2). These lesions include high-grade dysplasia (HGD) for esophagus, adenoma high-grade dysplasia for colorectum, ductal carcinoma in situ (DCIS) for breast, pancreatic intraepithelial neoplasia (PanIN) for pancreas, prostatic intraepithelial neoplasia (PIN) for prostate, high-grade squamous intraepithelial lesions (HSIL)/ carcinoma-in-situ 3 (CIN3) for cervix, bladder carcinoma in situ/noninvasive papillary carcinoma for bladder, high-grade dysplasia for stomach, Bowen’s disease, actinic keratosis (AK), lentigo maligna for skin, squamous carcinoma in situ (CIS) for lung, oral epithelial dysplasia/ CIS for mouth, anal intraepithelial lesions (AIL) for anus, renal intraepithelial lesions (RIL) for kidney, serous tubular intraepithelial carcinoma for ovary. These lesions can be found in screening procedures of high-risk patients or using diagnostic biopsies in patients suspicious of cancer or in cases of incidental cancer. Precancers for hematologic conditions include Monoclonal B-cell Lymphocytosis (MBL) for Chronic Lymphocytic Leukemia (CLL) and Monoclonal Gammopathy of Undetermined Significance (MGUS)/Smoldering Myeloma (SMM) for myeloma. Although precancerous lesions are common in general population for certain types (example-10% in breast biopsies [33], 24% in colon polyps [34], 43% in pancreatic Intra- epithelial Neoplasia (PanIN1) [35], 49% of actinic keratosis (AK) in men [36], not all precancerous cells progress into cancer [37,38]. It was found that 20% of breast precancer [39], 15–25% of colon polyps [40], 10–15% of PanIN [41] and 10% of AK lesions [42] progress to invasive stages. Additionally, there is a long transition period from development of precancer to invasive cancer which can span from 5–15 years for colorectal [34], 8 years for breast [33], 25–30 years for skin [36] and more than 10 years for prostate cancers [43], with exceptions for lung (2 years) [44] and pancreatic cancers (5 years) [45]. However, the higher likelihood of progression from precancerous stage to aggressive form of cancer in many cases implies that it is important to have an accurate, objective, and easy-to use method/biomarker to identify critical initial events leading to aggressive premalignant lesions from normal. Although the histopathological progression of these lesions has been well characterized for many cancers, there is still a dearth of information on key molecular events leading the transition of precancerous lesions to aggressive, invasive cancer. New HTAN PCA initiative has started shedding light on comprehensive understanding of the molecular, cellular, and tissue alterations and the interactions of the various cell types that drive tumor development and progression, especially the progression from premalignant lesions to invasive cancer. For example, spatial precancer atlas of breast cancer revealed that when normal breast specimens with patient-matched DCIS and invasive breast cancer (IBC) were compared, distinct coordinated transitions in their tumor microenvironment (TME) were observed. Interestingly, myoepithelial disruption was more advanced in patients with DCIS that did not develop IBC, implying this process could be protective against recurrence by allowing immune infiltration [46,47]. Thus, the findings from HTAN PCA research efforts could have implication on improvements in risk stratification, early-detection, and development of cancer prevention/interception strategies. NCI’s 2023 Precancer Atlas initiative will address this need by developing comprehensive, dynamic, high-resolution, multidimensional, multiparametric, and scalable atlases of precancerous lesions and their surrounding microenvironment [48].

Figure 2:

Examples of Early Lesions and Transition to Invasive Cancer

Created by Biorender.com

Early Detection Biomarkers:

Early Detection Biomarkers refer to the genomic, epigenomic, proteomic, metabolic and metabolomic molecules derived from human specimens (blood, urine, saliva, tumor etc.) that inform risk and diagnosis for a disease/cancer [49]. Biomarkers of risk include hereditary cancers where the patients have germline pre-disposition to certain cancers. Some examples of such cancers/ hereditary cancer syndromes (with associated germline mutations in gene) include hereditary non-polyposis colon cancer (DNA mismatch repair genes- MLH1, MSH2, MSH6 or PMS2), hereditary breast and ovarian cancer (BRCA1, BRCA2), familial adenomatous polyposis (APC), Li-Fraumeni Syndrome (TP53), Cowden Syndrome (PTEN) and Von Hippel-Lindau disease (VHL). Diagnostic biomarkers include markers, which are either abnormal or/and are elevated in body fluids, which can provide information on the course of the disease. Table 1 lists commonly used diagnostic tumor markers for cancer that may have additional clinical utility in relation to disease management.

Table 1:

Common Use Early Detection Tumor Markers.

	Cancer Type	Specimen Type	Clinical Utility
Alpha-fetoprotein (AFP)	Liver cancer	Blood	Diagnosis, Treatment response
B-cell immunoglobulin gene rearrangement	B-cell lymphoma	Blood, bone marrow, or tumor tissue	Diagnosis, Treatment response, Check for recurrence
BCL2 gene rearrangement	Lymphomas, leukemias	Blood, bone marrow, or tumor tissue	Diagnosis, Treatment determination
Beta-human chorionic gonadotropin (Beta-hCG)	Choriocarcinoma and germ cell tumors	Urine or blood	Stage assessment, Prognosis, Treatment response
Bladder Tumor Antigen (BTA)	Bladder cancer and cancer of the kidney or ureter	Urine	Diagnosis, Check for recurrence
BCR-ABL fusion gene (Philadelphia chromosome)	Chronic myeloid leukemia, acute lymphoblastic leukemia, and acute myelogenous leukemia	Blood or bone marrow	Diagnosis, Treatment determination, Monitor disease status, Treatment response
BRCA1 and BRCA2	Breast, ovarian and cervical cancers	Blood or Saliva	Diagnosis, Treatment determination
C-kit/CD117	Gastrointestinal stromal tumor, mucosal melanoma,	Tumor, blood, or bone marrow	Diagnosis, Treatment determination
	acute myeloid leukemia, and mast cell disease
CA-125	Ovarian cancer	Blood	Diagnosis, Treatment response, Check for recurrence
Calcitonin	Medullary thyroid cancer	Blood	Diagnosis, Treatment response, Check for recurrence
CD19	B-cell lymphomas and leukemias	Blood and bone marrow	Diagnosis, Treatment determination
CD22	B-cell lymphomas and leukemias	Blood and bone marrow	Diagnosis, Treatment determination
Chromogranin A (CgA)	Neuroendocrine tumors	Blood	Diagnosis, Treatment response, Check for recurrence
Cyclin D1 (CCND1) gene rearrangement or expression	Lymphoma, myeloma	Tumor	Diagnosis
DNA mismatch repair genes (MSH2, MLH1, MSH6, PSM2)	Lynch syndrome (colon cancer)	Blood or Saliva	Diagnosis
Gastrin	Gastrin-producing tumor (gastrinoma)	Blood	Diagnosis, Treatment response, Check for recurrence
5-HIAA	Carcinoid tumors	Urine	Diagnosis and Disease monitoring
Human papilloma virus (HPV)	Cervical cancer	Cervical cells	Diagnosis
Immunoglobulins	Multiple myeloma and Waldenström macroglobulinemia	Blood and urine	Diagnosis, Treatment response, Check for recurrence
IRF4 gene rearrangement	Lymphoma	Tumor	Diagnosis
JAK2 gene mutation	Certain types of leukemia	Blood and bone marrow	Diagnosis
KIT gene (KIT)	Gastrointestinal Stromal Tumors	Tumor	Diagnosis, Prognosis
Microsatellite instability (MSI) and/or mismatch repair deficient (dMMR)	Colorectal cancer and other solid tumors	Tumor	Identify those at high risk of certain cancer-predisposing syndromes, Treatment determination
MYC gene expression	Lymphomas, leukemias	Tumor	Diagnosis, Treatment determination
MYD88 gene mutation	Lymphoma, Waldenström macroglobulinemia	Tumor	Diagnosis, Treatment determination
Myeloperoxidase (MPO)	Leukemia	Blood	Diagnosis
Neuron-specific enolase (NSE)	Small cell lung cancer and neuroblastoma	Blood	Diagnosis, Treatment response,
PCA3 mRNA	Prostate cancer	Urine (collected after digital rectal exam)	Diagnosis (repeat biopsy after negative biopsy)
PML/RARα fusion gene	Acute promyelocytic leukemia (APL)	Blood and bone marrow	Diagnosis, Treatment determination, Monitor disease status, Treatment response
Prostatic Acid Phosphatase (PAP)	Metastatic prostate cancer	Blood	Diagnosis for poorly differentiated carcinomas
Prostate-specific antigen (PSA)	Prostate cancer	Blood	Diagnosis, Treatment response, Check for recurrence
T-cell receptor gene rearrangement	T-cell lymphoma	Bone marrow, tissue, body fluid, blood	Diagnosis; Detection and evaluation of residual disease
Terminal transferase (TdT)	Leukemia, lymphoma	Tumor, blood	Diagnosis
Urine catecholamines: VMA and HVA	Neuroblastoma	Urine	Diagnosis
FoundationOne CDx (F1CDx) genomic test	Any solid tumor	Tumor, blood	Companion diagnostic test to determine treatment
Guardant360 CDx genomic test	Any solid tumor	Blood	Companion diagnostic test to determine treatment; General tumor mutation profiling
5-Protein signature (OVA1)	Ovarian cancer	Blood	To pre-operatively assess pelvic mass for suspected ovarian cancer

Modified and adapted from NCI source (https://www.cancer.gov/about-cancer/diagnosis-staging/diagnosis/tumor-markers-list

New Developments in Molecular Cancer Screening and Early Detection:

Germline Testing

Nearly 10% of all cancers are caused by inherited genetic changes/ germline mutations in cancer susceptibility genes [50]. Expanding NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines®): on recommendations of genetic testing to assess risk and management of hereditary cancers, including ovarian, breast, endometrial, colorectal cancers and pancreatic cancers reveals the increased importance of germline testing in cancer prevention, risk and treatment [51]. Currently, genetic testing involves screening for pathogenic/likely pathogenic (P/LP) germline variants in cancer predisposition genes including BRCA1, BRCA2, CDH1, PALB2, PTEN, and TP53 which are associated with increased risk of breast, colorectal, ovarian, pancreatic, and prostate cancer. It also provides recommended measures for genetic counseling/testing and management strategies for individuals carrying P/LP germline variants It has been found that germline mutations in BRCA1 and BRCA2, result in an increased risk for developing breast and ovarian cancer [52]. The patients carrying P/LP germline variants in BRCA1/ 2 have an increased risk of developing cancer at an early age and have risk of multiple primary cancers. Lynch syndrome is an autosomal dominant genetic disease caused by germline mutations in DNA mismatch repair (MMR) genes (MLH1, MSH2, MSH6, and PMS2). This cancer syndrome is associated with high risk of colon and endometrial cancers, accounting for about 1–3% of all colorectal cancer cases and 2–3% of all endometrial cancer cases [53].

Recently, for the first time the U.S. Food and Drug Administration granted marketing authorization to the Invitae Common Hereditary Cancers Panel, which is an in vitro next-generation sequencing based diagnostic test that can identify germline variants in 47 cancer predisposing genes associated with an elevated risk of developing certain types of cancer [54]. The Invitae panel includes BRCA1 and BRCA2, Lynch syndrome associated genes (MLH1, MSH2, MSH6, PMS2 and EPCAM), CDH1 (mainly associated with hereditary diffuse gastric cancer, and lobular breast cancer) and STK11 (linked with Peutz-Jeghers Syndrome). However, the challenges of incorporating germline testing in the clinical workflow are lack of clarity on the approach (targeted or universal gene panel testing), lack of knowledge of the clinical performance of the assays in racially diverse individuals and absence of standardized systems for implementation in the clinic.

Synthetic Biomarkers

Nanoparticles are a class of synthetic biomarkers that represent a novel approach in the field of medical diagnostics and personalized medicine. Nanoparticles are designed to specifically recognize and interact with molecular targets, such as proteins, nucleic acids, lipids, or metabolites. While traditional biomarkers are limited by dilution and signal-to-noise ratio, nanoparticles can be artificially designed to activate or enhance expression and track progression of precancerous lesions. One such method is attaching reporter molecules to particles larger than the filtration limit of kidneys (>8 nm). Otherwise undetectable tumor-secreted protease activity can be measured in urine samples after cleaving detectable markers from injected particles [55–57]. These particles can be developed into a multicancer biosensor by attaching multiple reporters that provide a complex overview of precancerous and cancerous progression and activity [58,59]. Synthetic cleavable reporters can also be attached to an enzyme substrate to form small molecule probes that directly target metabolic pathways [60], reactive oxygen species (ROS) [61], or volatile organic compounds (VOC) [62]. Gold nanoparticles (AuNP) have extensive uses in imaging technology due to their unique physical, chemical, and electronic properties. In addition AuNPs spontaneously accumulate a serum protein corona to reduce surface energy that can be used to characterize the circulating serum proteins [63]. These nanoparticle-enabled blood tests of PDAC patient plasma have shown promise in detection both in specificity and sensitivity [64]. The protein corona can be formed before application to stabilize molecular probes to the particles. Wu et al. utilized the protein corona of gold nanoclusters to stabilize fluorescent polystyrene nanoparticles conjugated to EpCAM aptamer to detect breast cancer cells in vitro and in vivo in mice [65]. Although the nanoparticle use in early detection seems promising, the risks and adverse effects for infusion of nanomaterials are unknown.

Quantum dots (QD) are a subset of nanoparticles consisting of semiconductor nanocrystals whose small size (2–10 nm) enable quantum mechanical properties. Within nanoscale semiconductors, a stimulus causes an electron to cross the band gap from the valence band to the higher energy conduction band; these QDs emit fluorescence when the electron returns to the valence band. The color produced is determined by the wavelength emitted by the band gap which itself is dependent on the size of the particle undergoing excitation [66]. Attaching molecular sensors (antibody, ligand, etc.) to QDs allows for studies in cancer biomarker detection, tumor identification, and tumor microenvironment mapping. However, quantum dots have been limited to mostly in vitro applications due to the potential for heavy metal ion toxicity [67,68], but recent efforts have been made to reduce toxicity through formulation of carbon [69] and graphene quantum dots [70]. Carbon dots attached to Tri aminoguanidine (TAG) as a receptor for citrate, a potential biomarker for prostate cancer [71], were able to detect citrate presence in both cell culture and human urine samples with low cytotoxicity [72]. Another challenge for in vivo QDs is the need for specificity and imaging that can penetrate to the depth of tumor sites. Liu et al. developed a nanosystem of near-infrared quantum dots equipped with the tumor-penetrating peptide iRGD for a mouse model of pancreatic and orthotopic breast tumors [73]. The QDs were designed to undergo signal quenching through cation exchange to reduce background signal and toxicity through release of metal ions that were cleared renally. Near-infrared imaging identified an accumulation of QDs in fibroblasts and extravascular tumor cells with high specificity. Although iRGD peptide offers promise for improving the in vivo tumor imaging in the preclinical models, this approach further warrants investigation with respect to specificity and uptake by non-target cells.

Genetically encoded synthetic biomarkers utilize genetic engineering techniques to integrate within targeted tumor cells or neighboring cells in the tumor microenvironment (TME) to detect, report, or respond to molecular signals indicative of cancer. Vector-based biomarkers utilize plasmids or viruses to insert secretable reporters directly into the tumor cells by targeting tumor-specific promoters [74]. An enhanced green fluorescent protein (eGFP)-encoded oncolytic herpes simplex virus was used in a human trial of peritoneal cytology in pancreatic cancer to detect micro-metastasis and correlated to patient outcomes [75]. Alternatively resident cells that target the TME can be engineered to seek a precancerous lesion or tumor and secrete biomarkers for detection. One such study used macrophages that were modified to release Gaussia luciferase (Gluc) when polarized to an M2 tumor associated macrophage (TAM) phenotype [76]. Alternatively, bacteria have emerged as a promising vector for genetic biomarkers [77,78]. In a mouse model of colorectal cancer, genetically modified bacteria (A. baylyi) have been programmed to leverage the bacterial horizontal gene transfer capability to develop drug resistance in response to uptake of the mutated KRAS gene shed from tumor cells. These bacteria samples were then transferred to an antibiotic selection agar plate to detect KRAS mutations [79].

Although, synthetic biomarkers hold a great promise in detecting cancer early, there are some caveats that needs to be addressed in the field including background noise due to off-target and on-target, off-tumor activation, lack of standardized preclinical models for precancer and early-stage cancer, and limited understanding of the biology of early lesions and their transitions into malignancy. Since the field is still in infancy, human testing of synthetic biomarkers in well-designed trials will ultimately determine its utility in early cancer detection.

Liquid Biopsy

The U.S. National Cancer Institute (NCI) definition of liquid biopsy is “A laboratory test done on a sample of blood, urine, or other body fluid to look for cancer cells from a tumor or small pieces of DNA, RNA, or other molecules released by tumor cells into a person’s body fluids. Liquid biopsy allows multiple samples to be taken over time, which may help doctors understand what kind of genetic or molecular changes are taking place in a tumor. A liquid biopsy may be used to help find cancer at an early stage. It may also be used to help plan treatment or to find out how well treatment is working or if cancer has come back [80].” Although tissue biopsy is long considered gold standard to diagnose cancer, it only provides a single snapshot of the tumor, suffers from selection bias and is not reflective of the inter- and intratumorally heterogeneity.

Additionally, multiple biopsies from the tumor are sometimes performed, this process is limited in relation to the chance of targeting/ accessibility of tumor, surgical complications, and associated costs. On the other hand, liquid biopsy is relatively easy to perform, does not have selection bias for the tumor region as opposed to a tumor biopsy and can be used for longitudinal monitoring. Over the past two decades liquid biopsy sources have expanded to include various biofluid specimens including blood, urine, cerebrospinal fluid, saliva, amniotic fluid, ascitic fluid examining cancer-derived circulating tumor cells (CTCs), circulating nucleic acids including cell-free DNA (cfDNA), circulating tumor DNA (ctDNA), cell-free RNA (cfRNA) including mRNA, long non-coding RNAs (lncRNAs) and microRNA (miRNA), extracellular vesicles (EVs), tumor-informed platelets, proteins, and metabolites (1, 2). Most of the FDA approved liquid biopsy tests are prognostic tests including Cell Search® Circulating Tumor Cell (CTC) Test, Cobas® EGFR Mutation Test v2, Guardant360® CDx, FoundationOne® Liquid CDx, and Therascreen® PIK3CA Mutation Analysis which aids clinicians to help identify patients with cancer who may benefit from specific FDA-approved treatments [81]. This implies that there is a critical need to develop diagnostic and molecular screening biomarkers in addition to the standard of care methods and to establish their clinical utility in context of precancer or cancer diagnosis and screening.

Microfluidic chips are devices that use small channels to manipulate and control miniscule amounts of fluid on the order of microliters or smaller to simulate the tissue microenvironment. The microchannels function by capillary action that allows the flow of fluids through the microchannels without requiring an external pressure source, but microfluidic platforms can also be powered by pumps or gravity as the design requires [82]. In the context of cancer detection, microfluidic technology offers innovative approaches that enable highly sensitive, rapid, and cost-effective methods for detecting cancer-specific biomarkers, circulating tumor cells (CTCs), and cfDNA from liquid biopsies. The SMILE (SAW-MIP Integrated Device for Oral Cancer Early Detection) platform was developed for point-of-care early screening of oral squamous cell carcinoma (OSCC) [83]. The system uses immobilized anti-EpCAM to selectively capture tumor cells before injecting a secondary solution for staining. Other microfluidic systems screen for precancerous and early stage OSCC via changes in the amount and function of salivary cytokines IL-6 [84] and IL-8 [85]. However, keeping the isolated CTCs affixed to the surface of the device when rinsing remains a challenge particularly in early cancer detection when few CTCs are in circulation. Microchannels lined with AuNP conjugated to anti-EpCAM irreversibly sequestered CTCs from human blood and improved sensitivity for cancer detection [86]. Cancerous cfDNA from liquid biopsies is heavily diluted with a short half-life therefore extraction requires precision and speed. A pressure and immiscibility based extraction microfluidic chip successfully monitored the progression of HER-2 type breast cancer from plasma cfDNA and recognized a point mutation in phosphatidylinositol-4,5-bisphosphate 3-kinase (PIK3CA) during liver metastasis [87]. To make this technology accessible for point-of-care screening, the device must be scalable to produce while the targeting antibodies must remain functional for extended periods of time. To overcome these restraints, new materials, rather than traditional PDMS, provide longer term capability and ease of fabrication [88,89]. Cryodesiccation of microfluidic devices preserves the bioactivity of the attached antibodies and substantially extends the storage time [90]. The implementation of microfluidic-based systems holds considerable potential for improving early diagnosis, monitoring disease progression, and providing cost-effective cancer detection. However, the technology has to be standardized and rigorously validated for efficiency, reproducibility, and potential clinical utility.

Some of these tests are single-organ marker tests for specific cancers like Epi proColon® and Bluestar Genomics’ 5hmC Assay while other assays are multi-cancer detection (MCD) assays like CancerSeek, Galleri®, and OverC MCD Assay. Currently, the only FDA premarket-approved liquid biopsy test being utilized for colorectal screening is the Epi proColon®. It is performed by detecting methylated Septin 9 (SEPT9) DNA using blood samples [91]. A multi-institutional study on 1544 patients across Europe and the USA found that Epi proColon® achieved a sensitivity of 68% and a specificity of 80% when comparing Epi proColon® to colonoscopies for screening of colorectal cancer [92]. The Bluestar Genomics’ 5hmC Assay is another test used for pancreatic cancer detection that received the Breakthrough Device designation in 2021. It is a blood-based test which measures the levels of the biomarker 5-hydroxymethylcytosine (5hmC). In a recently completed case-control validation study, the Bluestar Genomics test achieved sensitivity of 67% and specificity of 97% in a population of 2,150 patients over age 50 [93]. PCR screening for cell free EBV DNA in plasma samples for a cohort of 20,174 asymptomatic persons for early nasopharyngeal carcinoma showed positivity for 1.5% of the patients and 0.17% of the patients were found to have nasopharyngeal carcinoma on endoscopic evaluation [94].

Unlike conventional cancer screening methods that focus on specific types of cancer, MCD tests aim to detect various cancers simultaneously, including more than 60% of cancers that do not have screening in standard of care. Though MCD technology is intended to screen patients that are symptom-free, many of the published studies determining the sensitivity and specificity of MCDs assessed blood plasma from patients whose cancer diagnoses were known prior to the study [95–97]. CancerSEEK is one recently developed MCD based on circulating proteins and mutations in cfDNA [95]. The CancerSEEK study of 1005 patients with non-metastatic cancer showed test positivity in a median of 70% of eight cancer types (ovarian, liver, stomach, pancreatic, esophageal, colorectal, lung, and breast). Overall sensitivity ranged from 69–98% with a specificity of greater than 99% for the detection of five cancer types (ovary, liver, stomach, pancreas, and esophagus) in average-risk subjects. Further implementation of artificial intelligence is key to the development of MCDs but requires training data sets from established patient information [98]. The Cell-free Genome Atlas study was a case-controlled, observational study that integrated machine learning with methylation-based cfDNA to detect cancer signals across multiple cancer types with intermediate sensitivity [99]. Though known cancer status is important for validating sensitivity, specificity is required to make MCDs clinically effective. In a study of asymptomatic patients, DETECT-A was a prospective interventional clinical trial to evaluate an MCD blood test that evaluated 9,911 women with no history of cancer [100]. Positive tests were independently verified by positron emission tomography–computed tomography (PET-CT) scans determining a high specificity (> 99%) with one percent false positives and 65% of cancers were detected at an early stage, with sensitivity varying with the tumor type. In an ongoing prospective clinical trial, ASCEND is validating a classification algorithm for a new version of the CancerSEEK assay utilizing 1000 patients with diagnosed/suspected cancer and 2000 patients with unknown cancer status. Additional GRAIL MCD trials intended for use in screening asymptomatic populations include STRIVE, SUMMIT, PATHFINDER and the GRAIL/UK NHS partnership [101,102].

While determining cancer risk is the general goal of MCD, the greater diagnostic consideration is a tissue-of-origin (TOO) for the most likely organ for cancer development. Most of the TOO algorithms are proprietary and a black box. However, based on the published literature, circulating MicroRNA (miRNA) appears to be a prime candidate for TOO determination for MCDs [103,104]. Recent studies have taken a predictive model of miRNAnomics for stage 1 lung cancer and expanded to biliary tract, bladder, colorectal, esophageal, gastric, glioma, liver, pancreatic, and prostate cancers [105]. Matsuzaki et al. used a machine learning approach with miRNAnomics to predict TOO for 13 cancer types from 7931 serum samples with 88% accuracy overall and 90% accuracy for precancer through stage II [106].

The continued development of MCD tests represents a paradigm shift in cancer screening strategies by combining known biomarkers, high risk status of the patients, and artificial intelligence to determine cancer risk and potential sites of origin. In alignment with the Cancer Moonshot^SM, NCI launched a new research network, the Cancer Screening Research Network (CSRN), that will assess the effectiveness of MCDs to prevent cancer-related deaths. Under CSRN’s Vanguard study, NCI will begin enrolling 24,000 healthy people aged 45–70 in 2024 to lay the foundation for future studies [107].

Although MCD tests hold great promise to reduce cancer related mortality, there is limited information and evidence on the clinical benefit especially in context of early-stage disease. Most MCD technologies perform poorly to detect early-stage disease with sensitivity varying between 40 to 50% [108,109]. Another clinical challenge is lack of clarity on how to diagnostically resolve and clinically manage a positive MCD test. MCD tests are considered adjuncts to current screening modalities, which currently are not efficient for all cancer types. Blood-based tests, in combination with subject characteristics, have the potential to make current screening more efficient by providing personalized screening schedules to individuals based on their own risk. For other less common cancers for which population-level screening is not practical, blood-based tests could identify individuals at sufficiently high risk to be screened for these cancers. However, the current candidate MCD technologies require rigorous investigation with respect to safety and effectiveness for reducing the morbidity and mortality associated with the cancers. To address these issues, careful assessment of the clinical utility of the assays by clinical trials through a well-established infrastructure may be essential.

Future Directions:

The future holds promise for personalized risk assessment and inclusion of biomarkers that would have the potential to tailor screening according to risk, and for other cancers to identify individuals for close monitoring. Improved screening for these cancers and effective evaluation of liquid biopsy-based tests with high specificity for a broader set of cancers, particularly for those for which screening modalities are not available, have the potential to significantly reduce cancer mortality [110].

The process from the establishment of the diagnostic performance of a new biomarker to producing convincing evidence that population screening for the biomarker can reduce mortality may involve a sequence of well-organized phases: development of technically sound and systematically evaluated assay/s, promising performance of the assay/s in case vs control followed by prospectively screened population (intended-use population), screening in the intended-use population must lead to earlier diagnosis and, more treatable point in the disease’s natural history and lastly, the upgrading or stage shift should be translated into sustained mortality reduction in larger study cohorts [111]. The NCI’s Early Detection Research Network (EDRN) is addressing some of these needs by providing an infrastructure that is essential for developing and validating biomarkers for early cancer detection [112]. EDRN has established the phases of biomarker development to specify criteria for progression from one phase to the next [20]. The mission of the NCI’s EDRN is to discover, develop and validate blood- and tissue-based biomarkers and imaging methods to detect early-stage cancers and to translate these into clinical tests [112,113]. The EDRN is a highly collaborative program that consists of Biomarker Developmental Laboratories, Biomarker Reference Laboratories, Clinical Validation Centers, and a Data Management and Coordinating Center. EDRN has successfully completed several multicenter validation studies. Additionally, the EDRN has built biospecimen, including organ specific, reference sets and informatics resources that are available to investigators both within and outside the EDRN.

In addition, artificial intelligence (AI) methodologies including computational disease modeling have emerged as a successful tool for risk stratification and early cancer detection [114]. Many companies, academic centers, and government agencies are attempting to build databases to use data for risk prediction, identify patients for clinical trials, and develop drugs [115]. However, AI in general healthcare faces several challenges including ethical considerations, governance and data access and security. Therefore, developing an ethically compliant framework for AI use in early cancer detection will be an important step. Since research on biomarkers is becoming more digitized and a vast amount of data is being collected through high dimensional omics and imaging technologies, sharing of such data has become paramount in leveraging methodologies and resources across scientific disciplines.

Overall, the early detection assay/s should demonstrate safety, accuracy, acceptability, cost-effectiveness and reach in the general populations. Once these tests have been shown to reduce death from cancer, they need to be integrated in the healthcare system. Early detection should be accessible to all populations, particularly the groups with disproportionately high cancer mortality rates and should not lead to overdiagnosis and overtreatment. With the advancement in novel early detection technologies together with enhanced knowledge of cancer biology, we have an opportunity to translate these in early cancer detection to potentially reduce the cancer associated morbidity and mortality.

Key Points:

1. Screening and early detection for clinically significant cancer is key to prevention.

2. Defining and characterizing the precancer lesion/state provides an opportunity for early cancer prevention, and interception.

3. New developments in molecular cancer screening and early detection including well designed MCD trials will pave way for cancer prevention and reducing the cancer screening and detection related mortality, morbidity, and associated cost.

Synopsis:

Cancer remains to be one the leading causes of death worldwide, primarily due to the late detection of the disease. Cancers which are detected at an early stage may enable more effective intervention of the disease. However, most cancers lack well-established screening procedures except for cancers with an established early asymptomatic phase and clinically validated screening tests. Additionally, despite the technological advancements in early detection biomarker field, there is still lack of in-depth knowledge on precancerous lesion/state for many cancer types for identifying markers of aggressive disease at an early stage. Further, there are unique challenges in establishing the utility of assays in terms of prioritizing and incorporating these biomarkers in the clinical management of patients. Therefore, there is a critical need to identify and develop assays/tools in conjunction with imaging approaches for precise screening and detection of aggressive disease at an early stage. New developments in molecular cancer screening and early detection include germline testing, synthetic biomarkers, and liquid biopsy approaches including multi-cancer detection tests.