The first preclinical tele-robotic models were first developed in the late 90’s and beginning of the twenty-first century, which were aimed at different clinical applications [11,12,13].

One of the first models to be commercially available for clinical applications was the MELODY system in France. It was the first model to use a mock US probe with a 3-DOF manipulator but required an assistant at the patient site to perform the compression and sliding of the probe [14, 15].

Since then, different models have been commercialized for clinical use, improving the overall design such as a higher DOF (up to 6–7) to reduce the intervention of the assistant. Other improvements included diminishing the latency between the doctor’s and patient’s end as well as better audiovisual communication and real-time US parameters adjustment. The history, development, and future directions of the tele-robotic US are beyond the scope of this article. We refer to two published reviews that deepen in this manner [5, 16].

Since the commercialization of the first clinical models there have been several studies assessing the feasibility of the tele-robotic US, generally using the standard US as a reference. There have been studies conducted in abdominal, obstetric, trauma/musculoskeletal, echocardiography, vascular and even thyroid imaging [8, 9, 14, 17,18,19,20,21,22,23,24,25].

The first published studies in abdominal imaging were done with the MELODY system and its precursors. Several studies in the first decade of the twenty-first century assessed different abdominal organs and structures such as the liver, portal vein, gallbladder, pancreas, spleen, kidneys, bladder, aorta and urogenital structures [17,18,19,20].

The visualization rate of the reported structures was generally above 80%, although some studies reported a lower visualization rate in some organs such as the spleen [17] and pancreas [18].

In 2017, Adams et al. evaluated the MELODY system in abdominal imaging in a cohort of 18 patients. 92% of the abdominal organs were correctly visualized with the tele-robotic US and it was also one of the first to assess quantitative measurements such as kidneys sagittal length and spleen long axis among others. There were no statistically significant differences when measuring liver parameters, spleen and proximal aorta but showed differences when assessing both kidneys in sagittal length, common bile duct diameter and distal aorta diameter [14].

Later, Adams et al. (2022) also conducted another study with the MELODY system in a remote area with 82 patients and a total of 87 explorations, of which 35 were abdominal. Among all the studies, 28% were considered inadequate for diagnosis and the rest adequate or adequate with reservations [21].

The era of 5G technology has allowed the development of new clinical studies with improved tele-robotic devices. Duan et al. (2021) evaluated the use of tele-robotic US in in the Intensive Care Unit in a total of 33 patients and assessed several diagnostic and incidental findings in the abdominal organs including the liver, gallbladder, pancreas, spleen, kidney as well as examinations for pleural and abdominal effusion. There was a moderate consistency between both techniques with a Kappa value of 0.6 [9]. Zhang et al. (2024), found no significant differences when measuring the aorta, portal vein, gallbladder, and kidney (longitudinal diameter). However, tele-robotic US underestimated the transverse diameter of the kidney [22].

Our results are consistent with the reported literature when assessing abdominal structures. There was an overall moderate/substantial correlation when assessing the left hepatic lobe and short and long axis of both kidneys in the post-trained cohort. There were also no statistical differences when assessing all the quantitative and qualitative variables in the post-trained cohort except the visual assessment of short axis of the left kidney and the measurement of the long axis of such kidney. In the latter, the difference between both means was approximately 2 mm with similar standard deviation, so it is not clinically relevant. Moreover, most of the incidental findings were also captured with the tele-robotic system including liver steatosis, gallstones or hydronephrosis. A few small hepatic and renal simple cysts were missed by the tele-robotic US. This is in line with other published studies [23,24,25].

Based on the results and the differences between the pre-trained and post-trained cohorts, the liver (especially the right lobe) and kidneys are the hardest structures to evaluate with the tele-robotic US. The location of these structures demands the reposition of the transducer in specific angles or planes in comparison to other structures such as the aorta, pancreas or urinary bladder that are easily accessible in supine position in a standard plane. The limited movement of the robotic arm in comparison to the hand, although being excellent, can contribute to these organs being the hardest to evaluate. Also, the lack of left-side video-camera limits the visualization of left-side structures like the kidneys. In such cases, the portable video-camera was allocated at the left side of the patient.

Another important factor to be considered is the presence of a marked subcostal liver, which limits the assessment of this organ through the subcostal plane. In these cases, the liver is better assessed through the intercostal plane but, in our experience, it is hard to master the visualization of this organ through this plane with the tele-robotic US. We hypothesize that a limited visualization and perspective of the ribs may contribute to this limitation.

Some other limitations found when evaluating these structures were increased body habitus, abundant bowel gas or presence of Chilaiditi syndrome, as also noted in other studies [25].

It is yet to be seen if more experience and training is required to halve these limitations in evaluating these structures.

Patient collaboration is also paramount for a good assessment since participants were frequently asked to move into a specific position or take a deep breath to better assess some structures such as the kidneys. As noted, restrictions in arm movement that complicates reaching some locations might contribute to this dependency. A study made with patients with COVID-19 pneumonia noted that the poor cooperation of the patients was one of the factors that difficulted performing the tele-robotic US [23].

We did not evaluate the use of tele-robotic US in patients seeking urgent medical attention. The fact that some structures such as the gallbladder and the kidneys were correctly assessed with this technique could highlight its role emergency radiology, especially in remote areas or patients that need to be isolated. Duan et al. (2021) found good correlation in a cohort of patients in the Intensive Care Unit when assessing gallbladder wall thickening, enlarged gallbladder, cholecystitis or bilateral hydronephrosis. Only two cases of hepatic cysts and one gallbladder polyp were missed by the telerobotic US, which per se are not relevant findings in patients with acute abdomen [9].

The caliber of the common bile duct was not assessed in our study because it was not considered to be feasible in non-emergency patients. Moreover, the higher resolution of the standard US might allow the visualization of the non-dilated common bile duct whereas this will not be achieved with the tele-robotic US and produce a bias. However, since the portal vein was assessable in almost all cases, it is presumably possible to detect a dilated common bile duct in the portal space.

Further studies are needed to better assess the role of tele-robotic US in patients admitted into the emergency room, as well as other clinical scenarios. Depending on the intended use, future tele-robotic US should be coupled with technically better US models.

The evaluation of the thyroid gland showed excellent concordance with the standard US, as shown in results. Zhang et al. (2022) did a comprehensive assessment of the clinical use of 5G-based tele-robotic US in thyroid disease in comparison to conventional US in a cohort of 139 patients. As in our case, there were not statistically significant differences between the measurements of the thyroid gland [8].

We also assessed thyroid nodules based on TI-RADS classification. All the nodules were correctly assessed except one TI-RADS 2 nodule which was missed by the tele-robotic US. Nodules were also correctly classified with the tele-robotic US except one TI-RADS 3 nodule that was misclassified as a TI-RADS 2. The tele-robotic US failed to detect calcifications in the cystic-solid nodule. Our results are in concordance with what Zhang et al. (2022) reported, in which only five of more than 100 nodules were missed by the tele-robotic US, and all of them were TI-RADS 3 [8].

The quasi-flat surface of the neck with easily accessible superficial structures allows for a good examination with the tele-robotic US. Lack of patient collaboration besides neck hyperextension and absence of conditions that may disturb the examination such as intestinal gas also facilitates the use of this technique. Moreover, thyroid disease is highly prevalent, and US is the gold-standard imaging modality for its assessment. In the light of the above, we believe thyroid examinations might be the first ones to be habitually performed by tele-robotic US, especially in remote areas.

To the best of our knowledge, this is the first study that an estimated learning curve for this technique is given. Based on our results the learning curve for achieving a good performance with tele-robotic US is fast and assessable. There was a decrease in the time needed to perform the US specially during the first 20 studies, and better overall results after that mark in abdominal explorations. In thyroid gland explorations there was an even faster learning curve, being almost excellent since the first patients.

We did not assess the learning curve in more than one radiologist. Ren et al. evaluated the variation trend of examination time in two sonographers and noted a downward trend despite the fluctuations for both. These results emphasize the progressive evaluation time reduction due improvement of experience and proficiency with the tele-robotic US [25].

The average time for performing the abdominal and thyroid gland explorations is in line with other studies that use a similar device, although there is a clear heterogeneity in the reported literature, probably because of differences in the protocol [9, 25].

Although not assessed in this study, the examination time is longer in tele-robotic US compared to standard US as demonstrated in the published literature [14, 22].

There was a high satisfaction/agreement rate when assessing the safeness and comfort of the tele-robotic US and communication with the radiologist. This is in line with other reported studies [8, 14, 21, 25].

Some patients reported excessive neck pressure during thyroid examination, a finding that has been previously described. Zhang et al. (2022) hypothesized that this may be due to the lack of adequate subcutaneous fat to reduce the force of US transducer and the cylindrical shape of the neck, which makes it harder for the probe to fit [8]. We believe the pressure on the neck should be also reduced to avoid this discomfort.

There was also indecisiveness about which technique is preferred. We believe that it will gradually gain more acceptance once it is implemented, as the performance is on par with the standard US. The possibility of reducing patient travel time might contribute to its acceptance, as noted in other studies [21].

A favorable cost–benefit is also important for the clinical implementation of tele-robotic US. There are few cost-analysis studies in which the feasibility of tele-robotic US has been assessed. Löfgren et al. found no clear cost differences at healthcare administration level if distant tele-robotic US was used instead of on-site traditional US for heart failure assessment with echocardiography. However, there was a reduced cost at personal level due to the patient-related costs (such as traveling) [26]. In a cost-minimization study, Adams et al. found that tele-robotic US and/or itinerant sonographer had a lower average cost than traveling radiologist/sonographer in rural and remote communities at healthcare and personal perspective [27]. Other studies have also shown affordable costs of this technique in rural areas [28]. More studies are needed to determine if tele-robotic US is economically viable in different scenarios besides rural and remote communities.

Patients, clinicians and healthcare providers acceptance are therefore paramount for the implementation of this technique in clinical scenarios, as well as appropriate regulatory policies [16]. Future studies will also determine the best scenario for tele-robotic US, whether it is in areas with low access to medical resources, emergency departments or other situations. Finally, the inception of artificial intelligence has the potential to strengthen the use of this technology. Autonomous robotic US might be able to scan several patients with minimal or without human intervention and detect relevant findings such as free fluid in polytraumatic patients. Also, it might be able to do standard examinations in outpatients which can be later tele-reviewed by a radiologist or specialist. This technology, however, is still under development [16].

There are several limitations in our study. First, the cohort in which the study was conducted had a normal mean BMI and patients were able to properly collaborate when performing the tele-robotic US, so the results may be overestimated for real-case scenarios. Future studies should aim at exploring the use of tele-robotic US in non-ideal scenarios such as patients with high BMI or impaired diseases. It is also especially important to assess its feasibility in patients seeking urgent medical attention in rural or remote communities due to the limited personnel resources in such places. Second, a single radiologist was responsible of conducting the tele-robotic US, not allowing for a comparison between different radiologists. Although relevant, the ability to correctly use the tele-robotic US also depends on the standard US experience of the subject. Third, the differences in the lower technical characteristics of the US model coupled with the robotic arm in comparison to the device used for the standard US can underestimate the tele-robotic results.

In conclusion, tele-robotic US has a good correlation with the standard US when evaluating abdominal structures, as well as a relatively fast initial learning curve and good overall acceptance. The performance when assessing the thyroid gland is even better, being almost identical to the standard US, which makes it a strong candidate for a future implementation. Despite these results, more clinical studies are needed to demonstrate its applicability in the daily clinical practice, especially in specific scenarios such as health checks-up or emergency departments.