Attitudes and Willingness of Cardiothoracic Group Physicians in the Cardiovascular and Radiology Departments toward the Adjuvant Use of CT-Derived Fractional Flow Reserve in the Diagnosis of Coronary Artery Disease

Study design and participants

This cross-sectional study was conducted in hospitals located in Beijing, Hebei, Shandong, and Guangdong regions from May to June 2023, with a focus on cardiothoracic physicians in cardiology and radiology as the primary study participants. The sample size was calculated using the Fisher et al. (23, 24) formula: n = Z² × P × (1-P)/D², where ‘n’ is the sample size, ‘Z’ is the Z-score corresponding to the desired confidence level (e.g., 1.96 for 95%), ‘P’ is the estimated prevalence of the event of interest (P = 0.5 was used here), and ‘D’ is the desired margin of error according to multiple sources (D = 0.072 here). Hence, the minimal sample size in this study was n = 196.

Inclusion criteria for participation were limited to cardiothoracic physicians in cardiology and radiology, while those with incomplete questionnaire responses were excluded. The distribution of questionnaires took place through WeChat groups utilizing the “Questionnaire Star” platform.

This study was approved by the Ethics Committee of Tianjin Medical University General Hospital (IRB2023-YX-173-01). All the participants who constituted the survey population provided written informed consent.

Procedures

The questionnaire design for this study was informed by previous publications (11, 12, 13, 14, 16, 17) and the “Chinese Expert Consensus on the Clinical Application of Coronary CT Fractional Flow Reserve” (18). Following the initial design, feedback was obtained from four experts, including two cardiovascular specialists (one associate chief physician and one chief physician) and two radiology specialists (both chief physicians). The questionnaire underwent revisions based on their valuable suggestions and subsequently underwent a small-scale pilot test. The reliability of the questionnaire was assessed, yielding a satisfactory Cronbach’s α coefficient of 0.884. The Kaiser-Meyer-Olkin (KMO) value was 0.828.

The final version of the questionnaire was developed in Chinese and comprised three dimensions:

1. demographic information, consisting of 10 questions;

2. attitude dimension, encompassing 15 questions, all rated on a five-point Likert scale ranging from 1 to 5 to gauge the participants’ attitudes, yielding a total score range of 15 to 75;

3. The willingness dimension, comprising eight questions, was also rated on a five-point Likert scale from 1 to 5 to assess the level of willingness, resulting in a total score range of 8 to 40.

The median of each score was used as the cutoff for determining poor or good KAP. The median is a statistically defensible and pragmatic cutoff in KAP studies, particularly useful when there is no widely accepted cutoff for “adequate” KAP, the score distribution is non-normal or contains outliers, and researchers need a context-specific threshold to categorize respondents for further analysis or intervention (25).

In order to facilitate data collection, an online questionnaire was created using the WeChat-based Questionnaire Star mini-program, and a QR code was generated so that participants could access and complete the questionnaire via WeChat. In order to ensure data quality and completeness, each IP address was restricted to one submission, and all items in the questionnaire were made mandatory. Data collected through the platform were exported to an Excel spreadsheet for further analysis. The research team members meticulously reviewed the questionnaires to ensure the integrity, internal coherence, and reasonability of the data.

Statistical analysis

Quantitative variables were presented as Mean ± SD. Intergroup comparisons for normally distributed data were conducted using the t-test or ANOVA, while the Mann-Whitney test or Kruskal-Wallis H test was used for non-normally distributed data. The attitude and willingness scores were not normally distributed, and the median was used as the cutoff point. Categorical variables were described in terms of frequency (percentage). The correlations between attitude and practice scores were examined using Spearman’s analysis. Univariable and multivariable logistic regression analyses were performed. Statistical significance was defined as a P-value less than 0.05. All statistical analyses were carried out using SPSS 26.0 software (IBM, Armonk, NY, USA).

Demographic characteristics

There were 275 questionnaires, but eight were excluded because the answer time was too short (<114 s) or too long (>1800 s), and two did not agree with their data being used for research, leaving 265 questionnaires in the analysis. The 265 cardiothoracic physicians in cardiology and radiology departments consisted of 151 males (57.0%) and 114 females (43.0%). Among the age groups, the highest representation was observed in the 31–40 years range, accounting for 45.3% of the participants, while those above 50 years were the smallest subgroup at 9.4%. A significant proportion of the physicians held a graduate degree (61.1%), followed by a bachelor’s degree (35.1%). The majority of participants worked as physicians (91.3%). Moreover, a considerable number of physicians had over ten years of work experience (54.7%) (Table S1).

Attitude and willingness

The correlation coefficient between the attitude and the willingness dimension was found to be 0.571 (p < 0.001), indicating a moderate positive correlation (Figure 1).

Figure 1

Many respondents reported being very familiar or familiar with CT-FFR (43.8%). Irrespective of familiarity, the respondents believed CR-FFR could avoid unnecessary invasive procedures (64.5%) and assess the entire coronary artery’s blood flow reserve (60.8%). Furthermore, about two-thirds agreed that CT-FFR could be helpful for post-treatment evaluation of coronary heart disease (67.9%), and two-fifths believed that its clinical application value could exceed that of invasive wire-based FFR (40.8%). However, some concerns were evident, especially about accuracy (38.1%) and potential for medical litigation (42.3%) (Tables 1 and S2).

About half of the participants strongly agreed or agreed (53.2%) that CT-FFR should become the standard method for diagnosing coronary artery disease. Moreover, 71.6% expressed willingness to try and apply CT-FFR technology if they had not used it before. Regarding patient education, 70.6% were willing to introduce the differences between CT-FFR and invasive examinations to the patients. About two-thirds of the participants expressed willingness to actively learn about the technical principles and clinical applications of CT-FFR and to receive further training and education on CT-FFR (Table 2).

Correlation analysis

The analysis revealed a strong positive association (r = 0.820, p < 0.001) between participants’ willingness to actively learn about CT-FFR technology (Q5) and their openness to receive further training (Q6). Additionally, participants’ willingness to actively learn (Q5) exhibited a moderate positive correlation with their interest in participating in clinical research involving CT-FFR (Q8) (r = 0.575, p < 0.001) (Figure 2).

Figure 2

Participants’ belief in the greater development potential of CT-FFR technology (Q12) showed strong positive correlations with their willingness to actively learn about CT-FFR (Q5) (r = 0.395, p < 0.001), their willingness to receive further training (Q6) (r = 0.453, p < 0.001), their consideration of the need for improvement in CT-FFR application (Q7) (r = 0.309, p < 0.001), and their willingness to participate in clinical research involving CT-FFR (Q8) (r = 0.487, p < 0.001). Similarly, participants’ perception of the necessity for further promotion of CT-FFR application (Q14) exhibited significant positive correlations with their willingness to actively learn (Q5) (r = 0.433, p < 0.001), their willingness to receive further training (Q6) (r = 0.482, p < 0.001), their consideration of the need for improvement (Q7) (r = 0.385, p < 0.001), and their willingness to participate in clinical research (Q8) (r = 0.489, p < 0.001). However, the rating of promotion of CT-FFR technology in the hospital (Q13) did not show significant correlations with the Willingness Dimension items (Figure 3).

Figure 3

Logistic regression

Among these factors, physicians were found to be less likely to have positive attitudes (OR = 0.312, 95% CI: 0.112–0.867, p = 0.026) compared with other occupations. Similarly, participants working in hospitals specialized in cardiovascular diseases showed a higher likelihood of positive attitudes (OR = 3.085, 95% CI: 1.224–7.771, p = 0.017) when compared to those in other types of medical institutions. After adjusting for potential confounding variables in the multivariable analysis, physicians (OR = 0.343, 95% CI: 0.123–0.957, p = 0.041) and participants in specialist hospitals (OR = 0.354, 95% CI: 0.140–0.896, p = 0.028) maintained significant associations with negative attitudes. The logistic regression analysis of participants’ willingness (categorized based on the cutoff value of ≥31 and <31) showed that a high attitude dimension score (OR = 6.280, 95%CI: 3.673–10.736, p < 0.001) was independently associated with willingness (Table S3).

Attitudes toward CT-FFR technology

The physician community shows a cautious optimism about CT-FFR, but there is a huge gap between this positive willingness and its practical application due to the multiple real obstacles like financial resources, workload, time, and access to the CT-FFR technology. In addition, concerns about the accuracy of CT-FFR can be related to motion artifacts, severe calcifications, and microvascular disease that may lead to differences between CT-FFR and invasive FFR (26). A significant number of participants displayed familiarity with CT-FFR and expressed confidence in its ability to reduce the necessity for invasive procedures while providing a comprehensive evaluation of coronary artery blood flow reserve. These attitudes are encouraging for the potential seamless integration of CT-FFR into clinical practice. The majority of participants held favorable views toward CT-FFR, with a notable percentage acknowledging its capacity to improve patient care by reducing the need for invasive procedures and facilitating a comprehensive assessment of coronary artery health. This cautious reception could be attributed to CT-FFR’s promising diagnostic performance and its potential to advance patient care in the field of cardiology (27, 28, 29).

Concerns among physicians toward CT-FFR technology

However, the study also sheds light on certain reservations among participants, particularly regarding the accuracy of CT-FFR results compared with invasive wire-based FFR. A notable proportion of participants expressed concerns about the potential risk of medical litigation arising from inaccurate CT-FFR results. Physicians are the ones signing the reports and bearing responsibility for misdiagnosis or inaccurate results. Physicians’ negative attitudes toward CT-FFR may be due to factors such as their professional awareness, sense of responsibility, education, industry culture, and traditional preferences. The lack of standardized interpretation is a major limitation of CT-FFR, which is recognized by the Chinese expert consensus on CT-FFR (18). These apprehensions signify that despite some positive attitudes toward CT-FFR, there exist lingering reservations and uncertainties regarding the technology’s reliability and potential legal implications. While this technique generally exhibits high diagnostic accuracy, specific cases of severe calcifications, poor imaging quality, or microvascular disease may lead to diagnostic inconsistencies. If not properly managed, these challenges can pose a risk of misdiagnosis or overtreatment. More education and training may be needed to promote physician acceptance of this new technology. In order to alleviate these concerns, studies should be conducted to standardize data interpretation protocols, improve algorithm robustness, enhance physician training, and enable liability insurance coverage as viable solutions. Additional studies demonstrating (or not) the effectiveness, safety, and healthcare economics of the technology are needed to reach an informed decision about its widespread use in the clinical setting. These measures would be designed to reduce the likelihood of controversy while increasing confidence in the clinical application of the technology. Consequently, it is imperative for the medical community and manufacturers to address these concerns diligently through persistent research efforts, validation studies, and transparent communication, elucidating the precision and safety of the technology. Manufacturers’ enthusiasm will have to be carefully balanced against a proper academic approach to ensure maximum profits for the patients and not for the industry.

Diagnostic accuracy of the CT-FFR technology

Previous studies have explored the impact of vessel calcification on the diagnostic accuracy of CT-FFR (29, 30, 31), further contributing to the understanding and refinement of this novel approach. Indeed, Tesche et al. (29) highlighted the potential for CT-FFR as a single examination that could guide patient management strategies. In addition, Tang et al. (30) showed that CT-FFR performed well in detecting lesion-specific ischemia. Tesche et al. (31) showed that the diagnostic performance of CT-FFR was related to the artery calcium scores. While CT-FFR has proven valuable in diagnosing and treating CAD, certain limitations persist. Clinical evidence derived from CT-FFR is contingent on specific software, cautioning against broad generalizations across different software products. In addition, CT-FFR demands high-quality CCTA images, leaving a substantial proportion of images unevaluable. The diagnostic efficacy of CT-FFR for severely calcified lesions remains uncertain, and a notable gap exists in clinical evidence supporting its application post-percutaneous coronary intervention or coronary artery bypass grafting. The presence of such concerns, specifically regarding the accuracy of CT-FFR results compared with invasive wire-based FFR and the potential for medical litigation, may impede the widespread implementation of CT-FFR (32, 33, 34).

Willingness to use CT-FFR technology

While the participants exhibited cautious attitudes, it is crucial to recognize the salience of accuracy-related concerns and the necessity of enhancing legal clarity to reach a better understanding of the pros and cons balance of CT-FFR technology. Regarding physicians’ inclination to acquire knowledge about CT-FFR, the findings unveiled that a majority of participants demonstrated receptiveness to CT-FFR. A majority of physicians displayed a willingness to explore and potentially apply CT-FFR technology, educate their patients about its advantages, and engage in enhancing their comprehension and proficiency in the technology. Furthermore, many indicated an eagerness to partake in clinical research endeavors involving CT-FFR. The study’s findings underscore a notable willingness among physicians to enhance their comprehension of CT-FFR technology, indicative of receptiveness toward integrating this innovative approach into their clinical practice. Furthermore, mounting evidence supports the utility of CT-FFR in guiding revascularization decisions, and the diagnostic performance of machine learning-based CT-FFR is anticipated to improve with advancements in image quality (10, 11, 12, 13, 14).

Cost-effectiveness of the CT-FFR technology

CT-FFR is an advanced noninvasive method to assess the physiological significance of coronary lesions. Despite its technical promise and guideline recommendations, its widespread clinical adoption remains slow, with one of the most formidable and systematic barriers being the lack of robust, generalizable cost-effectiveness data and unclear prospects for medical insurance reimbursement. Physicians are less likely to recommend or integrate a technology whose incremental clinical benefit versus cost is not convincingly established in the local healthcare context (35, 36). Mixed or preliminary cost-effectiveness evidence—such as data limited to certain populations or derived from short-term outcomes—feeds clinician uncertainty, especially when compared with well-established, reimbursed alternatives (e.g., stress tests, invasive angiography) (36, 37). When coverage or out-of-pocket costs are ambiguous, clinicians may avoid CT-FFR to shield patients from unexpected expenses and themselves from enterprising into nonstandard care pathways. This is particularly true since procedural and diagnostic costs can be substantial and variable (38). The financial policies of technology use trickle down to everyday practice. In settings with partial or delayed reimbursement trends—such as when some, but not all, payers or regions cover CT-FFR—clinical behaviors become inconsistent (37, 38). A lack of clear insurance endorsement means physicians cannot reliably predict patient access or institutional budget impact, further muting enthusiasm for broad adoption. The uncertainty around whether CT-FFR will be supported by insurance strongly correlates with lower stated intentions to use or trial the modality, even among clinicians otherwise convinced by the technology’s clinical promise. This is a classic finding in implementation science, and recent logistic regression analyses reinforce that reimbursement concerns are among the strongest predictors—often outweighing concerns about technical validity or workflow changes. Without clear cost-effectiveness, even hospital-level champions for innovation find it hard to justify the investment, as administrative and financial committees demand evidence of downstream savings or improved outcomes (35, 36, 39). Often, tertiary or quaternary centers with greater research focus and more flexible funding streams, specialist hospitals might pilot CT-FFR in select patient groups despite reimbursement uncertainty. However, even here, wider roll-out is often limited to research protocols or select populations as administrators are cautious to assume costs not guaranteed to be recouped (37). These settings, especially in resource-constrained or rural regions, are more sensitive to reimbursement clarity. They typically have less financial buffer for innovation risk and must prioritize technologies with clear, reimbursable pathways. Consequently, adoption of CT-FFR in these environments is highly restricted until reimbursement is formalized and/or local cost-savings are demonstrated in pragmatic terms, beyond trial settings (37, 40).

Correlations between attitudes and willingness

The correlation analysis between the attitude and willingness dimensions revealed a moderate positive correlation, indicating that physicians with higher attitude scores toward CT-FFR technology were also more inclined to engage in learning about it, seek further training, and participate in clinical research involving CT-FFR. This finding emphasizes the importance of addressing attitude-related factors when promoting the integration of novel technologies in clinical practice (41, 42). Comprehensive strategies need to be implemented to cultivate a proper attitude (i.e., balancing pros and cons) toward CT-FFR among cardiothoracic physicians in cardiology and radiology. Targeted educational programs could provide in-depth insights into the benefits and applications of CT-FFR, addressing any misconceptions or uncertainties. Interactive workshops could further facilitate hands-on experience and open dialogues, fostering a collaborative learning environment for physicians to share experiences and best practices. In addition, incorporating case studies that highlight successful outcomes with CT-FFR may serve to alleviate any skepticism and build confidence in its effectiveness. Furthermore, establishing ongoing communication channels and forums for continuous learning and feedback could ensure that physicians stay abreast of the latest developments about CT-FFR. Moreover, the correlation analysis demonstrated interesting associations among participants’ responses, indicating that higher attitude scores toward CT-FFR are closely linked to an increased willingness to learn, receive additional training, consider improvements, and engage in clinical research (43). This interdependence highlights the importance of fostering positive attitudes to stimulate active engagement with CT-FFR technology. Of course, attitudes and willingness are dynamic processes that can change, either positively or negatively, with accumulating data about new technology. Nevertheless, the findings emphasize the pivotal role of positive attitudes in influencing various aspects of willingness, reinforcing the need to cultivate proper perceptions of new technology. In China, the hospital classification system differs from that in Western countries; however, public tertiary hospitals in China are generally comparable to university medical centers or comprehensive medical centers in other countries. These hospitals play an important role in healthcare, providing high-level medical care, teaching, and research. The logistic regression analysis revealed that both physician status and employment in specialist hospitals were significantly associated with higher attitude scores toward CT-FFR, underscoring the potential influence of institutional and professional factors on technology acceptance (44). Furthermore, the study highlighted the critical role of attitude in motivating physicians’ willingness to embrace CT-derived FFR, emphasizing the need for interventions that enhance their understanding and appreciation of this innovative diagnostic method. These insights provide a valuable foundation for targeted strategies to enhance clinical practice, ultimately improving diagnostic accuracy and patient care. However, in contrast to expectations, the study did not observe statistically significant associations between demographic variables such as gender, age, and education level and the KAP scores toward CT-FFR.

Other technologies for hemodynamics

Strategies other than invasive angiography and noninvasive CCTA/CT-FFR exist to evaluate hemodynamics in CAD, including cardiac magnetic resonance imaging (CMR), radionuclide myocardial perfusion imaging, and stress echocardiography (17). Still, CT systems are usually more widely available than CMR systems; the examination takes less time and is less susceptible to motion artifacts, and the images have a higher resolution (17). CCTA does not involve radioactive isotopes that can have availability issues and pose certain safety risks. Echocardiography is operator-dependent, and accuracy or results can be variable among operators (45). Hence, CT-FFR appears promising in avoiding the limitations of the other modalities. Still, the availability of machines certified for CT-FFR can vary among countries and provinces, possibly affecting the willingness of the physicians to use them.

Integration of the CT-FFR technology in the clinical setting

A tiered training approach is recommended to improve the integration and effective use of CT-FFR technology among physicians. This approach should be tailored to the different levels of expertise and experience. For junior physicians and trainees, fundamental courses covering the basics of CT-FFR technology, including its principles, benefits, and limitations, should be implemented. These courses should emphasize hands-on training with simulation tools to build initial confidence. For mid-level physicians, intermediate training sessions focusing on case studies, interpretation of results, and integration of CT-FFR data into clinical decision-making are essential. Advanced workshops for senior physicians should delve into complex cases, troubleshooting, and the latest research findings.

In addition to these tiered training programs, regular interdisciplinary seminars should be organized to foster collaboration between cardiologists and radiologists. These seminars should highlight the clinical scenarios where CT-FFR is the most beneficial, discuss the accuracy and limitations of the technology, and review the current guidelines and best practices. Moreover, continuous professional development should include updates on the latest advancements in CT-FFR and feedback sessions to address any persistent challenges or concerns. Establishing a mentorship program where experienced users of CT-FFR guide less experienced colleagues can also enhance practical understanding and application.

Finally, transparent communication regarding the accuracy and limitations of CT-FFR, particularly concerning its use in specific clinical situations, should be emphasized. Clear guidelines on the interpretation of borderline CT-FFR values and their implications for clinical management should be developed to ensure consistent and accurate use.

Future perspectives

While CT-FFR has potential clinical benefits in CAD diagnosis (7, 8), its widespread adoption and application in clinical practice face challenges. CT-FFR is not currently included in practice guidelines in China despite available data that CT-FFR may improve decision-making, patient management, and resource utilization (10, 11, 12, 13, 14, 15). Still, there are some concerns about healthcare economics and the lack of standardized interpretation (16, 17, 18). Hence, performing additional studies and organizing expert opinion committees are essential to delineate the role of CT-FFR in the management of CAD and ultimately incorporate it in guidelines to guide its judicious use.

Limitations

This cross-sectional study examining the attitudes and willingness of cardiothoracic physicians in cardiology and radiology in China toward using CT-FFR as a diagnostic tool for CAD has several limitations. Firstly, the cross-sectional design restricts the establishment of causal relationships and temporal changes. Secondly, all participants were from a limited geographical area, limiting the generalizability of the results. Although some KAP studies are sometimes performed at the regional or national level, most KAP studies are performed locally because the differences in healthcare resources, healthcare policies, and training among regions can introduce variability in the results. Thirdly, potential biases, including sampling bias and social desirability bias (46, 47), may influence the validity of the results. The use of WeChat for sampling and questionnaire completion can lead to sampling and selection bias by excluding those with lower digital aptitudes and skewing inclusion toward those who are more active on social media. Fourthly, the questionnaire did not contain items about operational issues, and the processing time and workflow of CT-FFR were not considered in the present study and might affect the willingness of the physicians toward it, especially in the emergency setting. This study was a survey of the participants’ attitudes and willingness toward CCTA and CT-FFR, and the lack of knowledge and experience level of the surveyed population on CCTA and CT-FFR probably affected the results of this study to a certain extent. Nevertheless, the lack of knowledge of the population can also be further reflected in the attitude and willingness scores. Therefore, it is necessary to strengthen the publicity and education of CCTA and CT-FFR content and improve their recognition.