The role of imaging in health screening: screening for specific conditions

Abstract

There are well-established and emerging screening examinations aimed at identifying non-neoplastic and neoplastic conditions at early, treatable stages. The Radiology Research Alliance’s ‘Role of Imaging in Health Screening’ Task Force provides a comprehensive review of specific imaging-based screening examinations. This work reviews and serves as a reference for screening examinations for breast and colon cancer in a healthy population along with screening for lung cancer, hepatocellular carcinoma, and the use of whole body magnetic resonance imaging in at-risk individuals. American College of Radiology scoring systems, along with case-based examples, are included to illustrate the different disease entities. The future of screening is discussed, particularly in the context of artificial intelligence.

Introduction

Screening healthy and at-risk patients for malignancy and chronic diseases has emerged from large bodies of evidence validated in prospective cohorts for breast cancer screening, low-dose chest CT for lung cancer screening, CT colonography for colorectal cancer, and others. In cancer screening, many current imaging scoring and classification systems evolved from breast cancer with screening mammography and the BI-RADS scoring system (1–3). Additionally, screening for chronic but life-threatening diseases with imaging has roots in screening for abdominal aortic aneurysms with sonography, duplex examinations for carotid stenosis, and now includes calcium scoring with cardiac CT (4–6). Screening for malignancy or life-threatening non-neoplastic diseases is driven by perceived and evidence-based risk-benefits, reimbursement, published evidence, societal guidelines, and organizational expert statements. The following disease-specific sections provide an overview for screening for breast and colorectal cancer in healthy individuals, lung cancer in those with an appropriate smoking history, and hepatocellular carcinoma in at-risk patients. In addition, this paper reviews various indications of whole body MRI and screening for various benign diseases. Finally, we conclude the article by discussing the future of screening, particularly in the context of artificial intelligence.

Role of Imaging in Breast Cancer Screening

Apart from skin cancer, breast cancer is the most common cancer diagnosis and the second leading cause of cancer death in women. Approximately 12.4% of all women will be diagnosed with breast cancer at some point in their life (7). Screening mammography reduces breast cancer mortality by more than 40% in average-risk women aged 40 years and older (8). In addition to mortality reduction, early detection of breast cancer allows for a broader array of less invasive treatment options.

Mammography

Population studies demonstrate that the incidence of breast cancer increases with age. Therefore, more women in the younger age group (40–49 years old) will have to be screened for each life saved than in women aged 50 years or older. However, since women in a younger age group have a longer life expectancy, life-years gained for women diagnosed with breast cancer early because of mammography is higher than those in the older age group (9). There are various societies in the medical community with differing recommendations for when to start mammography for breast cancer screening in average-risk women (<15% lifetime risk), and how often to screen (Table 1). The major differences can be attributed to how each society views the risk to benefit ratio, with risks including false-positives and the possibility of over-diagnosis and benefits including mortality reduction and less invasive treatment options. Various organizations recommend starting screening mammography in average-risk women according to patient preference at ages 40–49 years while recommending starting screening for all women with average-risk at age 50 years. In their most recent guidelines, the American College of Obstetrics and Gynecology (ACOG) highlights that the decision about the age to begin mammography screening should be made through a shared decision-making process. Organizations also differ with their recommendations regarding the frequency of mammography (annual versus biennial). The American College of Radiology (ACR) and the Society of Breast Imaging (SBI) recommend annual screening beginning at 40 years of age and continuing while a woman’s life expectancy exceeds 5 to 7 years.

Table 1.

Breast cancer screening recommendations by different societies and organizations.

Women Age (years), Risk	U.S. Preventative Services Task Force (USPSTF), 2016 (4)	American College of Radiology/Society of Breast Imaging (ACR/SBI), 2017 (2, 5)	American Cancer Society (ACS), 2015 (6)	American College of Obstetricians and Gynecologists (ACOG), 2017 (7, 8)	American College of Physicians (ACP), 2015(9)
40–49, Average Risk*	Patient preference for when to start biennial screening	Annual screening	40–44 years=Patient preference for when to start annual screening 45–49 years=Annual screening	Patient preference for when to start annual or biennial screening	Patient preference for when to start biennial screening
50–74, Average Risk	Biennial screening	Annual screening	50–54 years=Annual screening 55–74 years=Biennial screening with choice to have annual screening	Annual or biennial screening	Biennial screening
≥ 75, Average Risk	Insufficient evidence for recommendation	Annually, while life expectancy > 5–7 years	Annually, while life expectancy ≥ 10 years	Patient preference for when to stop screening	No screening recommended
High Risk**	May benefit more than average-risk women from beginning screening in their 40’s	BRCA1/2, ≥20% lifetime risk = Annual adjunct breast MRI with mammography starting by age 30 years (but not before age of 25) Chest irradiation between 10–30 years of age= Adjunct annual breast MRI with mammography starting 8 years after treatment (but not before age of 25)	Annual adjunct MRI with mammography	Updated recommendations do not address.	Not addressed
Dense Breasts	Insufficient evidence for recommendation for adjunctive screening modality (i.e.: MRI, DBT)	Can consider adjunctive breast ultrasound	Insufficient evidence for recommendation for adjunctive screening breast MRI	Insufficient evidence for recommendation for adjunctive screening breast MRI	Not addressed

^*Average risk= <15% lifetime risk

^**High risk= BRCA 1/2 mutation (and their untested first-degree relatives), history of chest irradiation between 10–30 years of age, women with ≥20% lifetime risk of breast cancer, and other syndromes (including women and first-degree relatives with: Li-Fraumeni, Cowden, and Bannayan-Riley-Ruvalcaba syndrome)

High-risk women benefit from beginning screening mammography earlier than the average-risk population, in combination with supplemental screening (8). In BRCA 1 and BRCA 2 mutation carriers and other hereditary syndromes (such as: Li-Fraumeni, Cowden, and Bannayan-Riley-Ruvalcaba syndromes), screening mammography should begin by age 30 years but not before age 25 years. In women with mothers or sisters with premenopausal breast cancer, screening should begin by age 30 years but not before age 25 years or 10 years earlier than the affected relative at their time of diagnosis (whichever is later). In women with a history of chest irradiation between 10–30 years of age, screening should begin 8 years after treatment (but not before age 25 years. In women with a ≥20% lifetime risk of breast cancer, screening should start by age 30 years but not before age 25 years or 10 years earlier than the age of diagnosis of the youngest affected relative (whichever is later) (8–12). Justifications for the starting age of 25 years (but not before) include low incidence of BRCA1- or BRCA2-related breast cancer in high-risk women younger than 25 years and that women younger than 25 are often not included in randomized trials (13–19). One analysis of 22 studies pooled data from 6,965 female breast cancer cases to generate risk estimates for different age groups with BRCA1 or BRCA2 mutations (19). The authors found that annual breast cancer risk was very low in the age 20–24 group (0.02% annual risk for both BRCA1 and BRCA2). Compared to the annual breast cancer incidence in the 20–24 years age group, this increases 5-fold in the 25–29 years age group, and 18- to 36-fold in the 30–34 years age group (19).

The sensitivity of standard two-dimensional mammography is dependent upon breast tissue density, with sensitivity decreasing as density increases. However, three-dimensional tomosynthesis can address some of these limitations with the creation of thin-section images through the breast, which can decrease the lesion-masking effect of overlapping breast tissue. A case-based example in demonstrated in Figure 1.

Figure 1.

Screen-detected BI-RADS 5 breast mass. 55-year-old woman found to have a breast mass on screening mammography. Craniocaudal tomographic image of the left breast from a bilateral screening mammogram (A) demonstrates a mass (arrow) with calcifications in the medial left breast. Subsequent diagnostic mammogram (B) delineates the left breast mass on magnified craniocaudal view as spiculated with associated segmental and subsegmental amorphous microcalcifications which extend anteriorly towards the nipple. Ultrasound of the left breast shows a hypoechoic spiculated mass with posterior shadowing, corresponding to the mammographic abnormality (C). The diagnostic mammogram and ultrasound findings were in keeping with a breast mass highly suspicious for malignancy (BI-RADS 5). Post-contrast breast MRI demonstrates the spiculated left breast mass; susceptibility artifact from interval biopsy clip placement is noted along with non-mass enhancement extending towards the nipple. Ultrasound-guided biopsy and subsequent partial mastectomy demonstrated invasive ductal carcinoma.

Ultrasound

There is no data to support the use of ultrasound (US) for average-risk women with non-dense breasts (20). However, dense breast tissue lowers the sensitivity of mammography and increases breast cancer risk; therefore, according to the ACR and the SBI, for average-risk women with dense breast tissue, US may be useful as an adjunct to mammography for cancer screening (10, 21). The balance between increased cancer detection and the potentially increased risk of a false-positive examination should be considered in the decision to perform adjunctive screening US in these patients. Supplemental screening US is also indicated in high-risk women who cannot tolerate breast MRI (21, 22).

MRI

There is insufficient evidence to support the use of breast magnetic resonance imaging (MRI) for screening in average-risk women. However, in high-risk women, adjunctive MRI with mammography has the highest sensitivity (92.7%), and, therefore, the ACR, SBI, and the American Cancer Society (ACS) recommend adjunctive MRI with mammography in these patients (23–27).

Role of Imaging in Colorectal Cancer Screening

Colorectal cancer (CRC) is the second leading cause of cancer related deaths in the United States and is most commonly diagnosed in sixth and seventh decades of life. Research by the US Preventative Services Task Force (USPSTF) suggests that CRC screening is beneficial for average risk (4–5% lifetime risk), asymptomatic adults between the ages of 50 and 75 years (A recommendation; grade A category is recommended by the USPSTF with a high certainty that the net benefit is substantial). Screening in patients ages 76 to 85 years should be based on individual factors such as life expectancy, overall health, comorbidities and screening status (C recommendation; grade C category is recommended selectively by the USPSTF based on professional judgment and patient preferences with at least moderate certainty that the net benefit is small) with most benefit in patients with no prior screening history. The USPSTF does not recommend CRC screening in adults 86 years or older as the risk and mortality from other causes outweigh the benefits (28). These guidelines do not apply to patients with predisposing factors for higher than average risk of CRC such as Lynch syndrome, familial adenomatous polyposis syndrome, inflammatory bowel disease or a prior history of adenomatous polyps or previous colorectal cancer (28, 29). Many professional organizations encourage initiation of screening with increased frequency at a younger age with colonoscopy in patients with a positive family history of CRC or any other predisposing factors. Screening can prevent the development of CRC by removal of precursor lesions (polyps) and can improve outcomes by detecting asymptomatic early-stage disease (28–31).

To the best of our knowledge, there are no head-to-head comparative studies regarding the effectiveness of various screening strategies. Hence, engaging patients in the decision-making process regarding the best form or combination of screening methods can lead to higher levels of adherence over time. The risks associated with CRC screening result mainly from optical colonoscopy, including complications from sedation, bleeding, along with rare complications, such as postpolypectomy coagulation syndrome and colonic perforation. Risks also potentially includes work-up association from any incidental findings on imaging-based screening examinations (28, 29).

Stool based tests such as FOBT and FIT should be utilized annually for screening, can be used at home and do not require bowel preparation or anesthesia. The detection of intact human hemoglobin in stool increases the sensitivity of FITs as compared to guaiac-based FOBTs. The use of multi-targeted stool DNA tests is recommended every 1–3 years, which combines a FIT with detection of altered DNA biomarkers in cells (Cologuard®) shed into the stool to offer increased single-test sensitivity as compared with FIT alone at the expense of decreased specificity (31, 32). Another in-vitro screening test, the plasma SEPT9 gene methylation test (Epi proColon®), is currently the only commercially available United States Food and Drug Administration (FDA) approved blood-test for colorectal screening in the US (31).

Optical colonoscopy is recommended every 10 years and serves as the standard test for CRC screening. Colonoscopy also allows for concurrent intervention (polypectomy and biopsy) if necessary. Disadvantages include bowel preparation, anesthesia use, transportation requirements after the procedure and complications such as perforation, bleeding and required additional tests for incomplete evaluations. Alternatively, flexible sigmoidoscopy is recommended every 5 years and has similar but decreased complication rates as compared to complete colonoscopy (28, 30).

Imaging-based options play an important role as the primary or supplemental modality for colorectal cancer screening. Traditionally, double contrast (DCBE) was utilized primarily as a low-risk screening option for detection of colonic polyps or narrowing intraluminal lesions. Single contrast barium enema (SCBE) is a less sensitive option in patients who cannot tolerate insufflation. The ACS recommends DCBE use every five years in the list of options for CRC screening (29). However, the emergence of diagnostic CT has drastically reduced the popularity of ACBE in clinical practice. Decreased utilization of barium enema has negatively impacted training and proficiency in performing and interpretation of these studies. Hence, historical data regarding barium enema as a screening test likely does not accurately reflect current practice (29).

The USPSTF recommends performing CT colonography (CTC), also known as virtual colonoscopy, every 5 years as a non-invasive alternative to optical colonoscopy for CRC screening (29, 36). Limitations for the use of CTC include radiation exposure and the quality of the bowel preparation, including fecal tagging (36). Maximum radiation exposure in each CTC exam is ~7 mSv. In comparison, the average background radiation exposure in US is 3 mSv per year per person (29). Adherence to ACR practice parameters and reporting guidelines improve test performance and reduces reporting variability (29, 35). ACR also maintains a registry that provides evidence-based health outcomes and logistical data. This data allows participating facilities to compare results to the regional and national benchmarks for quality improvement (36). Data from the USPSTF assessing the performance of CTC without bowel preparation in the detection of polyps >10 mm found sensitivity of 67–90% and specificity of 85–97% (28). Positive CTC findings lead to confirmatory colonoscopy, which then subsequently dictates further follow-up management. CTC can also be utilized as an intermediary test in patients with positive stool-based tests. A meta-analysis showed that CTC has sensitivity of 88.8% (95% CI, 83.6%–92.5%) and specificity of 75.4% (95% CI, 58.6%–86.8%) for adenomas ≥6mm or colorectal cancer in 622 patients with positive FOBT results (29). A case-based example in demonstrated in Figure 2.

Figure 2.

CT colonography performed in a 72-year old woman with a positive fecal occult blood test. She was a poor candidate for sedation due to multiple medical comorbidities. 3D rendered luminal image (A) and source left lateral decubitus CT colonography image (B) demonstrate a fungating mass in the cecum (arrows) in keeping with a C-RADS C4 interpretation. Subsequent optical colonoscopy image (C) further delineates the cecal mass (arrow). Pathology at right hemicolectomy demonstrated a large villous adenoma.

Lastly, CTC can also be used as a complementary test in patients’ who receive inadequate or incomplete colonoscopy. In a study of 546 patients who underwent CTC after incomplete colonoscopy, 13% of patients had ≥6mm polyps. CTC after incomplete colonoscopy detected colorectal cancer in 9% of patients (33). Ideally, dedicated CTC bowel preparation on a later date after incomplete colonoscopy results in much higher exam quality compared with the same-day CTC. Patient should also ingest a fecal tagging agent (usually a small amount of oral iodinated contrast) after recovery from sedation at least two hours before the CTC is performed if the same day CTC is elected (34, 36).

Management of incidental findings on CTC remains an important challenge. Current literature suggests that incidental findings occur in 40–70% of CTC cases but require definitive treatment in only less than 5% of cases (28, 36). This discrepancy suggests that vast majority of incidental findings are clinically insignificant, which may add unnecessary work-up cost and anxiety. However, CTC can also provide ancillary benefit of concomitant screening for abdominal aortic aneurysm (AAA), osteoporosis and non-colonic malignancy (36).

There are many options for CRC screening; each with specific advantages, disadvantages and limitations. Visualization-based options include optical colonoscopy, CT colonography (CTC), flexible sigmoidoscopy and barium enema (BE). These exams require pretest bowel preparation for stool cleansing. Stool-based tests include fecal occult blood test (FOBT), fecal immunochemical test (FIT) and multi-targeted stool DNA test. Despite the wide spread availability of these tests, nearly one-third of eligible adults in the United States do not receive any form of CRC screening. The majority of patients with CRC are diagnosed at the middle or late stages of the disease (31). In summary, CRC screening can significantly reduce the mortality and morbidity from one of the leading types of cancer in the world. Imaging modalities, specifically CTC, play a vital role in detection of pre-cancerous lesions and early asymptomatic disease.

Screening for Lung Cancer

Lung cancer is the second most common cancer with an estimated 234,030 new cases in the US in 2018 (37). It is the leading cause of cancer death, with an estimated 154,050 deaths each year in the US in 2018 (37). Globally, there were an estimated 2.1 million new cases of lung cancer and 1.8 million deaths from lung cancer in 2018 alone (38). Smoking is the major risk factor for lung cancer with other risks being radon gas, occupational/environmental exposure (second-hand smoking, asbestos, metals, chemicals, radiation, air pollution, diesel exhaust), tuberculosis, and genetic susceptibility (37). Lung cancer is more amenable for curative treatment at earlier stages, either with lobectomy in combination with chemotherapy and/or radiation for non-small cancers or with a combination of chemotherapy and radiation for small call cancers (37). Early lung cancers account for only 16% of all lung cancer diagnosis (37), most of which are discovered incidentally in an imaging test. Unfortunately, the lung cancer is in an advanced stage in those who have clinical symptoms. Treatment options at this stage are usually palliative, including chemotherapy, targeted drugs, immunotherapy and radiation. The overall 5-year survival of lung cancer is only 18%, but this improves to 56% in the early stage localized cancers (37). The high prevalence, well-known risk factors, poor outcomes, long pre-clinical phase in some cancers and curative treatment options for early stages make lung cancer suitable for screening with imaging.

Chest radiograph (CXR) and low-dose CT (LDCT) scan are the imaging options for screening lung cancer. The National Lung Cancer Screening Trial (NLST), a multi-center randomized controlled study done on 53,454 asymptomatic patients at high-risk of lung cancer, showed that LDCT screening reduced lung-cancer mortality by 20% and overall mortality by 6.7% compared to those who were screened using a posteroanterior CXR only. With LDCT screening, 70% of cancers were diagnosed at earlier stages, compared to only 34% in an unscreened population (39). The Dutch-Belgian NELSON randomized controlled trial performed on 15,789 high-risk people using LDCT resulted in 58% of cancers being diagnosed at earlier stages (stage IA or IB) with significant mortality benefits observed over a period of 10 years (40).

The results from NLST led to establishment of lung cancer screening programs in multiple institutions with a multidisciplinary team approach. Eligibility criteria for lung cancer screening has been developed by different societies, with minor variations (Table 2). Lung cancer screening requires a physician visit, during which eligibility is evaluated and the patient is counselled on smoking cessation as well as adherence to the screening program. A collaborative shared decision-making approach between the patient and the physician is followed with a risk-benefit analysis. Risk-prediction models and decision aids may be used for this purpose (41). The obvious benefits of screening include reduced morbidity and mortality from lung cancer and its treatment. Ancillary benefits include decreased anxiety with a negative test, access to smoking cessation programs and a decline in smoking, especially after a positive scan (42). Risks include overdiagnosis, false positivity, incidental findings requiring more tests, radiation dose, psychological distress and the effects on quality of life. The radiation dose is low, typically less than 1.5 mSv, not exceeding 3 mSv, which is lower than a routine chest CT by more than 80% (42, 43).

Table 2.

Recommendations for LDCT lung cancer screening from different societies.

Society	Year (in years)	Age	Smoking
USPTF	2013	55–80	≥ 30 pack years Current or stopped smoking< 15 years
CMS	2013	55–77	≥ 30 pack years Current or stopped smoking< 15 years
NCCN	2015	55–74	≥ 30 pack years Current or stopped smoking< 15 years
		>50	≥ 20 pack years and one additional risk factor (Occupational exposure, radon exposure, history of cancer, family history of lung cancer, history of lung disease)
ACS, ALA, ACCP, ASCO	2013	55–74	≥ 30 pack years Current or stopped smoking< 15 years
AATS	2013	55–79	≥ 30 pack years
		≥50	≥ 20 pack years and one additional risk factor (Occupational exposure, environmental exposure, chronic obstructive pulmonary disease, prior cancer or thoracic radiation, family history, genetic factors)

Abbreviations

USPTF- United States Preventive Services Task force

CMS- Center for Medicare and Medicaid Services

NCCN- National Comprehensive Care network

ACS- American Cancer Society

ALA- American Lung Association

ACCP- American College of Chest Physicians

ASCO- American Society of Clinical Oncology

AATS- American Association for Thoracic Surgery

Overdiagnosis (i.e., detection of disease that does not affect mortality) is seen in 18% of screened patients. However, this is lower than that of breast (54%) and prostate (44%) cancer screening programs (43). A positive scan is a lung nodule > 4 mm without a benign calcification pattern. Of the 24% LDCT studies that are interpreted as positive, only 4% are true positives, while 96 % are false positives. False-positive results provoke anxiety in patients, which takes time to improve (44). The majority (96.4%) of the positive cases end up needing additional imaging, with 8.4% undergoing an invasive procedure (45). A negative screen does not necessarily exclude lung cancer, since there is a chance for interval cancers, which are typically more aggressive. Patients with comorbid conditions, who are unlikely or unwilling to complete curative treatment and with reduced life expectancy (< 10 years) are not recommended screening. If the patient is eligible for lung cancer screening, written order for LDCT is provided. Ineligible patients can still have a regular chest CT with contrast and smoking cessation programs can be offered.

CMS provides eligibility requirements for establishing a lung screening program, including requirements for the radiologist, imaging facility, and specific CT acquisition and dose criteria scanning technique and radiation doses (46). Nodule detection can be improved by using a maximum intensity projection reconstruction and/or computer aided detection (CAD) (47). Nodules can be measured either using diameter (uni-dimensional, bi-dimensional or average) or volume and volume doubling time, which was used in the NELSON trial (40). LDCT is reported using a standardized Lung CT Screening reporting and Data system (Lung-RADS) lexicon (Table 3) to effectively communicate the results, avoid confusion in interpretation and monitor outcomes (48). There are specific management guidelines for each category, with additional modifiers for clinically significant findings not related to lung cancer or history of prior lung cancer diagnosis. Generally, benign categories require annual 12 month follow-up LDCT, whereas the higher categories require additional investigation. The classification was updated in 2019 with a Lung-RADS 1.1 (49) (Table 3). The updated version has new size criteria for non-solid nodules, added peri-fissural nodules < 10 mm to category 2, revised mean nodule diameter measurement to measure long and short axis to one decimal point and report the mean average diameter to one decimal point, added volumetric measurements, assigned category 4B management for new large nodules and eliminated the C-modifier category (49). The use of structured reports helps in standardized documentation of the imaging techniques and the pertinent findings as well as embedding management guidelines. A case-based example in demonstrated in Figure 3.

Table 3.

Lung RADS system, adapted from references 41 and 42.

Category	Descriptor (prevalence)	Findings	Management
O	Incomplete (1%)	Prior chest CT located for comparison or all of lungs cannot be evaluated	Additional lung cancer screening CT images and/or comparison to prior chest CT
1	Definitely benign (90%)	No nodules Benign type of calcifications	Continue annual screening with LDCT in 12 months
2	Benign appearance of behavior	Solid nodule < 6 mm; new < 4 mm Part solid nodule < 6 mm Non solid nodule < 30 mm (20 mm on Lung Rads 1.0) or ≥ 30 mm and unchanged/slowly growing Category 3 or 4 nodules, unchanged for ≥ 3 months Peri-fissural nodules < 10 mm	Continue annual screening with LDCT in 12 months
3	Probably benign (5%)	Solid nodule ≥ 6 to < 8 mm at baseline or new 4 to < 6 mm Part solid nodule ≥ 6 mm with solid component < 6 mm or new < 6 mm Non solid nodules (ground glass) ≥ 30 mm (20 mm in Lung-RADS 1.0) on baseline CT or new	6 month LDCT
4A	Suspicious (2%)	Solid nodule ≥ 8 to < 15 mm at baseline, or growing < 8 mm, or new 6 to < 8 mm Part solid nodule ≥ 6 mm with solid component ≥ 6 mm to < 8 mm or with a new or growing < 4 mm solid component Endobronchial nodule	3 month LDCT PET/CT if ≥ 8 mm solid component
4B	Suspicious (2%)	Solid nodule- ≥ 15 mm or new or growing and ≥ 8 mm Part solid nodule with a solid component ≥ 8 mm or a new or growing ≥ 4mm solid component	Chest CT with or without contrast, PET/CT and/or tissue sampling PET/CT if ≥ 8 mm solid component Lung-RADS 1.1- For new ≥ 8 mm nodule in annual repeat screening CT, a 1-month LDCT for excluding infectious/inflammatory lesion
4X	Suspicious (2%)	Category 3 or 4 with additional or imaging features that increase the suspicion of malignancy	Chest CT with or without contrast, PET/CT and/or tissue sampling PET/CT if ≥ 8 mm solid component
S	Clinically significant non lung cancer finding (10%)	Modifier- Can add to any category 0–4	As appropriate to the specific clinical finding
C (In Lung-RADS 1.0; eliminated in 1.1)	Prior lung cancer diagnosis	Modifier- Can add to any category 0–4	-

Figure 3.

Lung cancer screening CT (A) in a 71-year old male patient shows a 1.4 cm spiculated nodule (arrow) in the right lower lobe, which is suspicious for malignancy. B. PET/CT done in the same patient shows high uptake in the nodule (arrow). Biopsy showed adenocarcinoma.

The lung cancer screening program should be robust to follow-up the patients after screening to ensure that they are managed appropriately. Smoking cessation programs should be available for current smokers. For benign categories, written orders are given for follow-up CT. If screening is positive, patients are referred to a specialized center. Screening ceases when patients are no longer eligible because of older age, longer duration since stopping smoking or development of comorbidities. Data from the scan is collected and submitted to a CMS-approved registry. Auditing and quality control should be performed on all aspects of the program, including scanning, reporting, counselling, shared decision making and follow-up.

Following USPSTF and CMS reports, lung cancer screening has been widely adapted by multiple institutions (50). Extrapolating the results of NSLT to the US population indicates that potentially 8.6 million people would have met the criteria from screening and that potentially 12,000 deaths from lung cancer were prevented (51). Absence of infrastructure and insufficient personnel were reported to be the most common barriers for implementation (41). Using the Lung-RADS criteria in a clinical screening program reduced false positivity by up to 75%, especially in the lower-risk nodules (45) and in a clinical screening program, reduced the overall positive rate from 27.6% to 10.6%, and increased the positive predictive value 2.5 times to 17.3%, without an increase in false negative results (52). Compared with no screening, CT screening for lung cancer from the NSLT showed an incremental cost-effectiveness ratio of $52,000 per life-year gained and $ 81,000 per quality-adjusted life-year gained (52).

In future, the screening criteria may have to be reevaluated to include broader groups, since NLST enrolled only smokers and did not account for other risk factors as well as sex and ethnicity. Recent studies show that only 26.7% of individuals with a diagnosis of lung cancer met the NLST screening criteria (41). Organizations such as the National Comprehensive Cancer Network recommend expanding the eligibility for screening so that the underserved populations of different socioeconomic and racial backgrounds are included (Table 2) (53). Risk-prediction models can be used to determine screen eligibility for patients who fall outside the USPSTF guidelines (41).

Screening for HCC

Overview and Patient Population

Hepatocellular carcinoma (HCC) is the most common primary liver malignancy and it is the second leading cause of cancer death worldwide. Although HCC is not in the top ten causes of cancer death in the US, it is the most rapidly growing cause of cancer death in the US (54). The primary risk factors for HCC are the presence of cirrhosis and hepatitis B or C virus. Approximately 80% of patients with HCC have hepatitis B or C virus and patients with cirrhosis have a 2 – 8% annual chance of being diagnosed with HCC (55, 56). Similar to lung cancer screening requiring an appropriate smoking history, screening individuals for HCC requires appropriate exposure history (chronic hepatitis B viral infection) or underlying liver disease (cirrhosis). The most cost-effective way to screen at-risk patients for the development of HCC is by ultrasound. The presence of a suspicious or equivocal finding on screening ultrasound may prompt more advanced imaging for diagnosis, including multiphase contrast-enhanced CT or MRI. Alternatively, contrast-enhanced ultrasound may be used. Patients may alternatively undergo primary screening with contrast-enhanced CT or MRI if quality of the sonographic images obtained during screening ultrasound is inadequate, such as a poor acoustic window due to patient habitus or hepatic steatosis.

LI-RADS

The ACR-sponsored Liver Imaging Reporting and Data System (LI-RADS) provides a standardized guideline for communicating liver imaging findings. Recent LI-RADS versions (including v2018) include a reporting system for screening abdominal ultrasound which communicates the benignity, worrisome, or equivocal observations. These categories include US-1 no reportable observations or only definitely benign observations, US-2 subthreshold an observation or observations <10 mm in size that are not definitely benign, and US-3 conferring a positive finding, US-2 may warrant a 3 to 6 month follow-up with ultrasound to assess for interval growth whereas a US-3 may warrant a diagnostic contrast-enhanced CT, MRI, or ultrasound. A separate component of US LI-RADS is the visualization score of the examination. Similar to reporting breast density classifications, the visualization score communicates if there were any technical limitations to the study including poor or incomplete visualization of the liver (55, 56).

LI-RADS: CT and MRI

Multiphase MRI or CT are modalities of choice to characterize indeterminate intrahepatic lesions or to screen certain patients who are poor candidates for ultrasound screening. MRI is typically the modality of choice. Multiphase CT has similar sensitivities in characterizing intrahepatic lesions compared to MRI and may be an ideal choice in patients with contraindications to MRI (e.g., noncompatible implants or devices), patients who cannot lie supine or hold their breath well for the duration of an MRI examination, and patients with contraindications to MRI contrast (e.g., allergy or patients on dialysis who are at-risk for nephrogenic systemic fibrosis but can receive iodinated contrast). Starting with version 2017 and now 2018, LI-RADS has a dedicated algorithm for liver imaging findings on CT and MRI, CT/MRI LI-RADS. These incorporate characteristic imaging features of HCC including washout, capsule formation, and arterial phase hyperenhancement. Additionally, there are categories of ancillary features that are categorized to include features that favor benignity, malignancy not specific to HCC, favor HCC, or favor a non-HCC malignancy (55, 56). A case-based example in demonstrated in Figure 4.

Figure 4.

Screening contrast-enhanced MRI in a 68-year-old man with hepatitis C virus-related cirrhosis. Arterial (A), venous (B), and delayed (C) T1-weight VIBE acquisitions demonstrate an approximate 3.5 cm mass (arrows) in segment 4A which demonstrate heterogeneous arterial enhancement, subsequent washout, along with a surrounding pseudo-capsule. This is in keeping with a LI-RADS 5 mass, representing hepatocellular carcinoma.

LI-RADS: Contrast-enhanced ultrasound

Contrast-enhanced ultrasound for characterization of indeterminate liver lesion has become more widespread and reported since the FDA approval of Lumason for liver ultrasonography in 2016. The algorithm for contrast-enhanced ultrasound in recent iterations of LI-RADS (versions 2017 and 2018) has the same overlapping imaging features for HCC as the CT/MRI algorithm, including arterial phase hyperenhancement, capsule formation, and the presence/absence of washout. Contrast-enhanced ultrasound (CEUS) may be a primary modality for imaging centers with experience and volume in CEUS, but its true advantage is in specific patient populations, such as patients with poor renal function but not on dialysis who cannot receive MRI contrast for risk of nephrogenic systemic fibrosis and are poor candidates for iodinated contrast for CT (e.g., advanced chronic kidney disease such as stage IV or V but are not on dialysis) (55, 56).

Implications for treatment

The definitive treatment of choice for HCC in the US is liver transplantation. HCCs are often first triaged with locoregional therapy or partial hepatectomy. Different iterations of LI-RADS have evolved over time to be incorporated into the American Association for Study of Liver Diseases guidelines. As advocated and endorsed by the Organ Procurement and Transplantation Network, HCC is one of few cancers that can be diagnosed without the need for biopsy if characteristic imaging features are present (55, 56). LI-RADS categorizations have been correlated with surrogates of overall survival independent of histopathologic diagnosis. Choi et al. (57) reported a retrospective cohort of 194 patients with a heterogeneous group of HCC and non-HCC malignancies (intrahepatic cholangiocarcinoma and combined intrahepatic cholangiocarcinoma-HCC). Choi et al. (57) found that LI-RADS categorization was an independent predictor of overall and recurrence free survival and that LI-RADS category M had a poorer overall and recurrence-free survival compared to category 4 or 5. The sensitivity and specificity of LI-RADS category 5 for differentiating HCC from non-HCC malignancies were 69% and 87%, respectively; whereas the sensitivity and specificity of either LI-RADS category 4 or 5 was 86% and 62% (57). Differentiating HCC from non-HCC malignancy is pivotal for treatment algorithms, as non-HCC malignancies may affect one’s eligibility for transplantation.

Whole Body MRI Screening

Since its introduction into clinical use, MRI has been performed to assess individual organs (i.e. brain or spinal cord) or a specific anatomical area. Historically, scanning with a larger field of view (FOV) and/or covering multiple anatomic stations in one exam was impeded by unacceptably long scan times. However, recent technological advancements, including multichannel techniques, high quality surface coil systems, parallel imaging and automatic table movement, have resulted in fast image acquisition and improved image quality and made it possible to image larger regions of interest with shorter scan times (58).

MR provides excellent soft tissue contrast and high spatial resolution and permits evaluation of the morphology, function and metabolism of tissue without radiation exposure. The lack of ionizing radiation makes whole body MRI particularly attractive for patients who require several imaging examinations throughout their lives, such as the evaluation of treatment response in oncologic patients (59, 60), imaging surveillance in patients with cancer predisposition syndromes (61) and evaluation of patients with chronic pathologies with multiorgan and multisite involvement. Some of the validated or emerging indications for whole body MRI include assessment of multifocal osseous, vascular, or neurologic diseases, seronegative arthritis (62), inflammatory diseases involving muscles or fascia (63), Langerhans cell histiocytosis (64), neurofibromatosis (65), and disseminated infection (63). Because the risks of ionizing radiation are greater in pediatric patients, imaging protocols using MRI are preferred over CT or nuclear medicine (66, 67). This is of particular importance in children with Li-Fraumeni syndrome (Figure 5), a hereditary cancer syndrome due to mutations in the TP53 tumor suppressor gene, as this condition has been linked to radiation-induced cancers in addition to many other malignancies (68). Utilization of whole body MRI for surveillance in pediatric cancer predisposition syndromes was first highlighted in 2011 by Monsalve et. al (69), who recommended annual whole body MRI for patients with Li-Fraumeni syndrome. The benefits of MRI surveillance in Beckwith-Wiedemann syndrome, hereditary retinoblastoma survivors and DICER1 syndrome (61, 70, 72) have also been reported.

Figure 5.

Whole-body MRI in a 6-year-old male with Li-Fraumeni syndrome and a history of right thigh embryonal rhabdomyosarcoma status post resection and chemotherapy. (a) Single station coronal T1-weighted contrast enhanced image shows the lesion at baseline (arrow). Coronal head-to-toe T2 HASTE (b) and single station coronal turbo inversion recovery magnitude (TRIM) (c) images demonstrate post-treatment changes secondary to prior right vastus intermedius and vastus medialis rhabdomyosarcoma resection with mild atrophy of the right thigh musculature and no recurrent or new disease.

There is limited research comparing the performance of whole body MRI at 1.5T and 3T field strength. Whole body MRI protocols typically image the body in four to six anatomic stations with multiple body coils and multiple imaging planes per station. Pulse sequence selection is highly variable at different institutions. According to published whole body MRI protocols for pediatric oncology, coronal short-tau inversion recovery (STIR), axial half-Fourier single-shot turbo spin echo (HASTE) sequences, and diffusion-weighted imaging (DWI) are the most commonly used sequences (72). Coronal STIR efficiently evaluates a larger field-of-view than axial sequences and is particularly useful for identifying osseous lesions (73). It is prudent to have a variety of different pulse sequences available which can be tailored to best address the individual patient’s clinical scenario. Whole body MRI can be further complemented by targeted small field-of-view imaging for problem solving. Also, given the low sensitivity of MRI in the detection and characterization of pulmonary findings, in selected cases, these findings can be further assessed with chest CT (74).

Contrast-enhanced sequences can be useful for specific indications such as improved detection and characterization of lung, liver, bone and brain lesions (75), but may result in longer examination times (58, 76). Some institutions use iron oxide nanoparticles such as ferumoxytol as an alternative to gadolinium given recent studies reporting gadolinium deposition in the brain (77, 78). However, ferumoxytol is not yet FDA-approved for pediatric applications. The use of DWI may potentially obviate the need for contrast-enhanced imaging and the risks associated with contrast administration (76).

Routine whole body MRI scan time averages between 30 to 60 minutes from localization to completion (58, 74). The total scan length is determined by a combination of different factors including the size of the FOV, choice and number of sequences, the use of intravenous contrast material, patient cooperation, and MRI technologist skill and comfort level with the exam. Although referred to as whole body MRI, the extent of body coverage on a whole body MRI is variable in different patients and institutions, mainly as a result of incomplete coverage of the lower extremities (74). The field-of-view in routine whole body MRI usually includes the neck, chest, abdomen, pelvis and lower extremities at least through the level of the knees (74). Brain MRI is not performed as part of the routine whole body MRI in some centers. A case-based example in demonstrated in Figure 5.

Having a universal definition of the extent of body coverage that constitutes a whole body MRI is important for billing purposes. Currently, no specific Current Procedural Terminology (CPT) code exists to describe whole body MRI (70). It has been suggested that screening whole body MRI be reported using an unlisted MRI CPT code which could necessitate prior authorization that might hinder prompt patient access (79). Further research is needed to standardize whole body MRI technical parameters, lexicon, reporting and billing practices as well as management of incidental findings discovered with whole body MRI.

Screening for Non-Malignant Diseases

Calcium Screening

Myocardial infarction (MI) is among the most significant causes of morbidity and mortality among adults in the United States (80). Tools predictive of coronary artery disease (CAD) and atherosclerosis may thus have utility in reducing the incidence of MI and improving outcomes in these patients. Traditionally employed metrics, such as the Framingham Risk Score as well as the Pooled Cohort Equations have been successfully used to predict the risk of CAD and MI (81). Over the years, calcium scoring has emerged as an important tool in determining the presence of CAD as well as in the risk stratification of these patients. Cross-sectional images of cardiac CT scans are used to gauge the amount of calcium present in the coronary arteries. The cumulative calcium score has been consistently demonstrated to be a reliable tool in predicting acute coronary events. In a systematic review and meta-analysis of four studies, Pletcher et al. reported greater coronary artery calcium scores to be independent predictors of coronary heart disease events (82). Similarly, Sharma et al. (83) highlighted that the coronary calcium score offered additional benefit to the Framingham Risk Score in predicting MI as well as in differentiating ischemic from non-ischemic cardiomyopathy. Arad et al. (84) reported that the coronary score was a more accurate predictor of acute events than standard risk factors and C-reactive protein levels. As patients with coronary scores of zero have been shown to have a very low risk of subsequent MI, several studies have noted the potential utility of the coronary artery calcium score as a screening test in the acute setting (85, 86). Despite this, however, there is a dearth of evidence from well-designed trials demonstrating improved clinical outcomes in patients who underwent coronary artery calcium screening compared to those who did not. As such, the USPSTF has concluded that there are insufficient data to warrant a positive recommendation for testing at this time (87). Further studies, including comparisons between coronary score and other imaging tests, such as myocardial perfusion imaging (MPI), as well as additional level I evidence regarding the utility of calcium screening can provide further data regarding the role of this important tool in patient care. A case example is demonstrated in Figure 6a.

Figure 6.

Screening for non-malignant diseases. A) CT images quantifying mild coronary artery calcifications in a 62-year-old-woman. B) Abdominal ultrasound image show a 4.1 cm abdominal aortic aneurysm in a 67-year-old man with a substantial smoking history undergoing screening for abdominal aortic aneurysm.

Abdominal Aortic Aneurysm Screening

Abdominal aortic aneurysms (AAAs) represent a dilation of the abdominal aorta and are relatively frequent in the geriatric population, with the prevalence in men older than 50 years exceeding 7% (88). Although smaller aneurysms and even larger, unruptured aneurysms are often asymptomatic, aneurysmal rupture can result in catastrophic hemorrhage and consequently high mortality rates approaching 90% (89). By contrast, the mortality in patients undergoing elective surgery for AAAs ranged from 5 to 7% (90, 91). As such, identifying AAAs prior to rupture has been hypothesized to have the potential to improve patient outcomes by providing additional treatment options, such as endovascular stent-graft placement, on an elective basis. At present, abdominal ultrasound is the preferred imaging modality, being readily available, avoiding radiation exposure, and demonstrating strong accuracy, with sensitivity and specificity rates approaching 100% (90).

There are numerous studies demonstrating a significant relationship between screening for AAAs and improved patient outcomes. Wanhainen et al. (92) reported data from a randomized trial where 253,896 men were screened and subsequently underwent endovascular or open surgical repair of their aneurysms. The authors concluded that screening was not only associated with reduced mortality, but that surgical intervention resulted in very low rates of perioperative mortality and that screening was cost-effective (92). Reduced rates of mortality following screening have been corroborated in numerous additional randomized controlled trials (93–95). However, one study by Norman et al. reported that population-based screening for AAAs was not associated with reduced overall mortality (96). Given the level of supporting evidence, the USPSTF has recommended screening for AAAs in men over the age of 65 years who had a history of smoking (97). The authors noted that the lower rates of AAA development in non-smokers may reduce the benefits of screening in this patient population. Despite this, the authors concluded that there is insufficient evidence to recommend screening in female smokers and recommended against screening females who had never smoked, citing increasing risk/benefit ratios with the increasing age of AAA development in these groups.

While screening via ultrasound facilitates the identification of aneurysms of various sizes, there remains considerable debate regarding the management of smaller AAAs (98). As such, additional studies examining various treatment approaches, including surveillance for small aneurysms in the context of AAA screening are warranted. A case example is demonstrated in Figure 6b.

Carotid Stenosis

Cerebrovascular accident (CVA) is among the leading causes of disability in the adult population. Carotid artery stenosis has been demonstrated to be a significant risk factor for CVA due to the potential for embolization to the cerebral vasculature (99). As such, the identification of stenosis and subsequent treatment via carotid endarterectomy (CEA), stenting, or intensive medical management has been thought to have the potential to reduce the incidence of CVAs. However, large studies have demonstrated that CVAs due to carotid artery stenosis are is relatively rare, with 45% of strokes in patients with asymptomatic carotid stenosis being due to cardioembolism or atherosclerosis of the deep subcortical vessels (100). Currently, the modality most frequently employed in screening for carotid stenosis is duplex ultrasonography of the vessels, which is relatively available and avoids radiation exposure, unlike other techniques such as CT angiography (CTA). However, several studies have demonstrated suboptimal sensitivity for minor stenosis and suboptimal specificity for significant stenosis when compared to CTA (101, 102).

At present, there are no published studies presenting level I evidence examining the utility of screening for carotid stenosis in asymptomatic patients. In addition, there is a dearth of tools that can be used to predict which patients presenting with carotid stenosis could benefit from revascularization. Existing randomized controlled trials have reported a slightly reduced incidence of CVA and death in patients undergoing surgery when compared to medical management; however, it is thought that the current conservative standard of care could result in an even smaller difference (103, 104). These limitations, coupled with the potential harms of operative intervention, have resulted in the USPSTF and the American Heart Association and the American Stroke Association to recommend against routine screening of asymptomatic patients (105). Future studies examining the role of screening in patients with carotid bruits, prior transient ischemic attack/CVA, and significant cardiovascular risk factors are warranted.

Value of screening for specific conditions

A comprehensive discussion of value in screening is beyond the scope of the current work and detailed in the companion Radiology Research Alliance ‘Role of Imaging in Health Screening’ Task Force’s companion manuscript on screening overview, rationale of screening and screening economics. As a summary and illustration of the value of different imaging screening examinations for specific conditions, cost per quality of life year is presented in Table 4 using referenced studies (52, 106–114), adjustment for inflation (115), and currency conversion (116).

Table 4.

Cost per quality of life year for different imaging screening examinations for various specific conditions. Cost from referenced data was adjusted for 2020 inflation using the U.S. Bureau of Labor Statistics Consumer Price Index Inflation Calculator (115). Cost was adjusted from the publication year of the referenced works, not the inclusive years from the references’ analyses.

Screening examination	Cost in USD per quality of life year (Cost USD adjusted for 2020 inflation)	Reference(s)
Annual screening mammography	$49,000 – $58,000 ($63,919 – $75,660)	106
CT lung cancer screening	$60,000 – $81,000 ($62,617 – $89,575)	107, 52
CT colonography	$19,416* – $45,222* ($20,263 – $53,985)	108, 109
Abdominal MRI for hepatocellular carcinoma	$25,202 – $68,212 ($25,899 – $72,661)	110, 111
Abdominal US for hepatocellular carcinoma	$64,072 ($68,251)	111
CT calcium scoring	$48,800 ($57,321)	112
Abdominal ultrasound for abdominal aortic aneurysm	$57,835* ($70,856)	113
Carotid Doppler	$39,495 ($63,221)	114

^*Data was converted to USD from authors’ original data using an Internal Revenue Service listed exchange rate (116) for the publication year.

Future of screening

Radiology is a technologically advancing field constantly assessing new imaging techniques and new modalities. Artificial intelligence (AI) and machine learning are current areas of interest in medicine and particularly in radiology and may have future roles in imaging-based screening examinations, such as semi-automated interpretation of screening mammograms, low-dose chest CTs for lung cancer screening, and CTC examinations (117). There has been some promising work with pulmonary nodule detection on CT using these advanced techniques (118, 119). Computer-aided detection has had a role in interpretation of mammograms for over two decades with a not as promising clinical performance (120). Due to marked improved performance of AI in recent years there has been a renewed interest in AI for interpreting imaging studies. AI algorithms with clinical performance comparable to breast radiologists could be used in different clinical scenarios such as triaging of normal screening mammograms that would not need human interpretation. AI may allow for machine-based identification of suspicious findings, following verification by human observers. This approach could result in a reduction in breast imagers’ screening workload which can allow more time for radiologists to spend on more complicated studies, potentially improving accuracy (121). McKinney et al. (122) published a high-profile study in Nature of an AI system used on a screening mammography dataset. The dataset contained over 25,000 screening mammography examination from two United Kingdom centers with a 1.6% rate of cancer and over 3,000 examinations from a single US center with a 22.2% cancer rate. The authors reported that the AI system outperformed the study radiologists in several metrics and categories including reductions in false positives and false negatives and a greater area under the receiver-operating characteristic curve. Although AI outperforming the study radiologists is noteworthy, there are several important limitations of the study. First, the study was based on screening mammography, not diagnostic mammography. Furthermore, the authors asked the readers to assign both a BI-RADS 0, 1, or 2 along with a BI-RADS score that did not include 0 (i.e., the readers had to rate it as BI-RADS 1, 2, 3, 4a, 4b, 4c, or 5). To the best of our understanding of McKinney et al.’s (122) data, their analyses were off the forced BIRADS 1 through 5 scale, not BI-RADS 0, 1, or 2 scale, which is inconsistent with clinical practice. The authors were vague about tomosynthesis, stating, “[some study radiologists] may have also had access to breast tomosynthesis images.” Cases were considered cancer negative if they had a follow-up screening mammogram interpreted as negative, which may confound the data with subtle developing findings identified by the study radiologists but labeled as cancer negative in the cohort. Lastly, it is unclear what proportion of their cohort had eventual cancers diagnosed by mammographic findings and a stereotactic biopsy versus indeterminant mammographic findings best seen with ultrasound and characterized with ultrasound-guided biopsy. Well-organized clinical trial will be needed to validate the findings by McKinney et al (122).

Delivery of imaging screening results to patients has opportunities for innovation. Wahab et al. (123) reported a cohort of women undergoing screening mammography who received their initial results either through a pre-recorded video message by a breast radiologist through email compared to traditional delivery methods (letter for BIRADS 1 or 2 and letter plus phone call for BIRADS 0). Messages were pre-recorded for both BIRADS 1 or 2 (‘normal’ [or benign] mammogram, repeat screen in one year) and BIRADS 0 (needing additional views). The majority of survey respondents in both groups (video and traditional methods) indicated that their preferred delivery of results would be through a video message from the radiologists (123).

There are developing uses and indications for imaging screening examinations. For instance, one trend to observe is the use of whole body MRI and its growing use in the adult population (125, 125). A large determinant will be if reimbursement for this examination is proportionate to the number of body parts imaged. Another growing trend with a growing body of evidence along with in-progress clinical trial is abbreviated screening breast MRI (126, 127). Future prospective studies are needed to assess the stand-alone performance of the proposed new AI algorithms in real life clinical settings. When adequately studied, these AI-based algorithms can provide access to screening imaging across the world even in areas where there is shortage of specialized radiologists and thus improve access to high-quality care for all (128). As AI and machine learning continue to develop, it is important that imaging-based screening remain relatively inexpensive. If costs go up or insurance coverage declines, then fewer people will undergo screening, and potentially treatable screen-detected cancers will manifest as larger, more dangerous cancer with a greater potential for metastatic disease and subsequent cancer-related mortality.

Conclusion

Imaging plays an increasingly central role in early detection of cancer and life threatening non-neoplastic conditions such as AAA in asymptomatic patients at-risk, with the promise of decreased morbidity and mortality from early intervention. Although controversies exist surrounding some imaging-based screening options and different organizations have slightly different screening recommendations, the overall mortality benefit of available screening examinations is widely acknowledged. Balancing potential benefits in mortality and morbidity reduction with potential harms of screening, including health hazards and economic costs at both the individual and the population level is challenging. Radiologists play an integral role in developing new techniques and optimizing imaging protocols to provide more effective screening tools and surveillance strategies that can positively influence patient health and contribute to better risk stratification and informed decision making.