Comparison between a handheld ultrasound device and a traditional ultrasound for performing transcranial sonography in patients with Parkinson's disease

Abstract Objective The aim of this study is to compare a portable ultrasound (US) device and a traditional US for performing transcranial ultrasonography (CCT) in patients with Parkinson's disease (PD). Methods This is a cross‐sectional, observational, and analytical study. The study recruited a total of 129 individuals from two public hospitals in the city of Rio de Janeiro in a prospective and non‐randomized manner between September 2019 and July 2021 as follows: group A with 31 patients with PD, group B with 65 patients with PD, and group C with 64 healthy individuals. Group A was used to collect data to establish the agreement analysis of the TCS measurements between the two devices. Groups B and C provided data for constructing the receiver operating characteristic curve for the handheld US. The subjects underwent the assessment of the transtemporal bone window (TW) quality, the mesencephalon area, the size of the third ventricle, and the substantia nigra (SN) hyperechogenicity area. Results There was a good agreement between the methods regarding the quality of the TW–Kappa concordance coefficient of 100% for the right TW and 83% for the left, the midbrain area—intraclass correlation coefficient (ICC) of 69%, the SN area ICC = 90% for the right SN and 93% for the left and the size of the third ventricle ICC = 96%. The cutoff point for the SN echogenic area in the handheld US was 0.20 cm2. Conclusions The handheld US is a viable imaging method for performing TCS because it shows good agreement with the measurements performed with traditional equipment, and the measurement of SN echogenic area for PD diagnosis presents good sensitivity and specificity.


INTRODUCTION
Transcranial sonography (TCS) is a neuroimaging technique used to obtain and evaluate images of the cerebral parenchyma in movement disorders through B-mode and assess the blood flow velocities of the intracranial vessels through Doppler mode Monaco et al., 2018). This is an ultrasonographic technique used to detect abnormalities in the echogenicity of substantia nigra (SN), thalamus, lenticular nuclei, red nuclei, the continuation of the median raphe nucleus, and ventricular diameters and assesses vascular reactivity Monaco et al., 2018;Walter et al., 2008).
The main limitation of TCS is the lack of bone window, which could vary between 5% and 44% of the population, according to Brisson, Santos et al. (2021), depending on various factors, including, sex, age, and ethnicity.
The B-mode on TCS as a tool to assist in the diagnosis of Parkinson's disease (PD) appeared in 1995 when Becker et al. (1995) noticed the increase in echogenicity of the SN for the first time in a group of patients with PD. Since then, the method has been the object of study of innumerable research studies, primarily in Europe, the Americas, Japan, and Korea, with the first publication in Brazil by Fernandes et al. (2011). Currently, the use of TCS as an auxiliary method in PD diagnosis using the B-mode scan is already well established.
TCS has numerous advantages over other neuroimaging methods.
As it is non-invasive, it does not emit ionizing radiation; it is performed in real-time; it has a short duration and relatively low cost, with no loss due to movement; and, above all, can be performed at the bedside as many and necessary times as possible (Ali et al., 2018;Berg et al., 2008). PD, the main cause of parkinsonism, is a slow, asymmetrical, idiopathic progressing disease, resulting in most of the progressive loss of neurons in the ventrolateral portion of the pars compacta of the SN. PD is characterized by the presence of bradykinesia and at least one of the two cardinal signs: rigidity and resting tremor, usually occurring in persons that are 50 years old or older. The diagnosis continues to be challenging when the motor signs appear, and at this point, complementary exams can help (Hughes et al., 2002;Postuma et al., 2015). SN hyperechogenicity is observed in about 90% of PD, with cutoff values between 0.20 and 0.25 cm 2 in most studies depending on the ultrasound (US) system used (Berg et al., 1999(Berg et al., , 2001. This image correlates with cellular iron and neuromelanin deposition and the microglia activation mechanism demonstrated in postmortem studies (Berg et al., 2010;Zecca et al., 2005). Almeida et al. (2022) in their research observed an estimated accuracy of 79.2% for the PD diagnosis by the TCS, while the 99mTc-TRODAT-1 DATSPECT was associated with an accuracy of 99%. This method has a significantly higher diagnostic accuracy for PD than TCS.
However, sonography does not use radiation ionizing or radiopharmaceuticals, has a lower cost, and is more available in health services, which makes this method an interesting tool.
With the advent of technological advances in recent decades, the US has become increasingly portable. More recently, new pocket US devices have become available, making it possible to use the handheld US, which brings ease of transport of the device to perform the exam (Ali et al., 2018). The handheld US has advantages over traditional equipment, the main one being portability as it is much smaller and lighter, which allows it to be easily transported to hard-to-reach places; another difference is the lower cost of its acquisition.
This research aimed to verify the agreement between a handheld US device and a traditional US for performing TCS in patients with PD.
Another research objective was to establish the SN hyperechogenicity area cutoff point for the handheld US. Considering this information, we did not find studies in the literature that evaluated the use of handheld US for performing TCS in patients with PD.

Study population
This is a cross-sectional, observational, and analytical study where Clinical and sociodemographic data were collected using a semistructured questionnaire.
The collected data from group A was used to establish the agreement analysis of the TCS measurements between the two devices: Philips Groups B and C provided data for the construction of the receiver operating characteristic (ROC) curve for the handheld US.
PD patients inclusion criteria in groups A and B: The PD patient's inclusion criteria are age over 18 years, with a minimum of 2 years of PD diagnosis and 5 years of follow-up. The diagnosis was performed following the MDS-PD 2015 criteria confirmed by two neurologists with more than 5 years of experience in movement disorder diseases.
PD patients exclusion criteria in groups A and B: Presence of atypical and secondary parkinsonism and previous neurosurgery.
Inclusion and exclusion criteria in group C: Age over 18 years of age and absence of neurodegenerative diseases, hyposmia or anosmia, previous stroke, and neurosurgery.

US examination techniques and TCS protocol
The US technique consists of placing the probe on the temporal region in front of the auricle of the ear, on the orbitomeatal line of the hemifacial region, and then searching for the temporal bone acoustic window, the thinnest region of the bone plate that allows the passage of low-frequency ultrasonic pulse.
The subjects were placed in a dorsal decubitus position and the examiner, on the right, in a supine position. To acquire images of structures of interest in the midline area bilaterally, the examiner then made circular movements with the probe over the TW.
The parameters for adjustment were the same on both equipment: frequency of 2 to 4 MHz, a dynamic range of 45 to 55 dB, and a penetration depth of the ultrasonic beam between 14 and 16 cm, contour amplification medium or high. Image brightness and time gain compensation were adjusted according to need. Both US systems had a phased array transducer (Koloudk et al., 2009;Walter & Skoloudik, 2014;Walter et al., 2007).
The subjects were evaluated for the quality of the TW, the mesencephalon area, the size of the third ventricle, and the SN hyperechogenicity area.
The criterion used for the insufficient window was the nonvisualization of the SN and midbrain. The insufficient bilateral window was an exclusion criterion for SN and midbrain analysis. However, in all cases, we evaluated the window quality parameter, and in some cases of bilateral insufficient window (for SN and midbrain), it was possible to visualize the third ventricle.
The measurements acquired were: the midbrain area in square centimeters on the axial plane, which, in the handheld US, was calculated using an ellipsis that allowed manual adjustment, which is the only method available to calculate area on this machine; and in the traditional US, it was manually circumscribed, the method available to calculate area on this equipment. The hemi-midbrain area ipsilateral was measured on both sets of equipment and multiplied by two to obtain the value of the total area of the midbrain.
Calculation of the SN hyperechogenicity was performed the same way as the measurement of the midbrain area on both sets of equipment. Hyperechogenicity of SN is defined by increased size, compared with normal ranges, and the basal cisterns were the referential used to identify the SN hyperechogenicity.
The largest transverse diameters of the third ventricle were measured. The size of the third ventricle was calculated in centimeters by measuring the distance between the linear hyperechoic inner layers of the ependyma, on the axial plane, ipsilateral, at the level of the thalamus (Walter et al., 2007). This technique was identical on both machines.
All the measurements were acquired after freezing and expansion of the images (Figure 1).
The reference value for the third ventricle according to Huber (2010) for individuals under 60 years of age is less than 0.7 cm and for individuals more than 60 years is less than 1.0 cm.The reference value for the entire midbrain area according to Aoun et al. (2021), by an experienced TCS rater, can have a mean value of 4.47 ± 0.53 cm 2 . The cutoff values of SN hyperechogenicity for PD can vary between 0.20 and 0.25 cm 2 depending on the US system used (Berg et al., 1999(Berg et al., , 2001. The exams were performed by two sonographers (with more than 5 years of experience with the method and more than 2 years with TCS).
In group A (sample of 31 patients with PD), each individual was evaluated at the same time by examiner 1 and then by examiner 2. Examiner 1 used traditional equipment, and examiner 2 used the handheld US.
One examiner did not have access to the other examiner's images or measurements. Groups B (all PD patients) and C (healthy individuals) were evaluated only by examiner 2 with the handheld US.

RESULTS
Among the 129 participants recruited, 65 presented with PD, and 64 were healthy control subjects. There was a predominance of males (57.36%) and the median age was 68 [58, 74] years old for parkinso- The mean duration of PD was 8 [6, 10] years. (Tables 1 and 2).

TCS with the handheld US for the ROC curve
The area under the ROC curve was 0.91 with a confidence interval of 0.95 (lower limit 0.84 andupper limit 0.97) The Youden index was 0.85 with the measure of 0.20 cm 2 representing the best cutoff pointfor the SN hyperechogenicity area to differentiate between sick (with PD) and healthy individuals in this sample (Table 3). The sensitivity of the test was 98.3% and specificity 86.7%. The positive predictive value (PPV) was 0.89, and the negative predictive value (NPV) was 0.9898%. Note: Distribution of concordance sample windows between systems in subjects with PD.

TCS for analysis of agreement between systems
The assessment of the right TW performed with the handheld US agreed 100% with traditional equipment. Of the 31 subjects from the A group examined with both equipment, it was observed that three subjects presented with insufficient right windows. The analysis of the inter-system agreement for the left TW reached a KCC of 83%.
Handheld US categorized 27 subjects with sufficient left TW, while traditional equipment described 28 subjects as sufficient (Tables 6   and 7).
The midbrain area was measured in 28 subjects from the PD group with sufficient TW. In the handheld US, the mean value of the measurement of the mesencephalon area was 4.8 cm 2 , while in traditional equipment, it was 4.7 cm 2 . The inter-system agreement was 69%, calculated by the ICC (Tables 8 and 9).
The third ventricle had its thickness checked in 31 subjects with permissible TW. In both systems, the values found were the mean and the median, 0.6 cm, and the ICC, 96% (Tables 8 and 9).
In 28 subjects from the PD group, where the RSN was assessed, both systems had mean and median values of 0.2 cm 2 and an ICC of 90%. The LSN was verified in 28 subjects from the PD group, where the handheld US measured a mean of 0.2 cm 2 and a median of 0.3 cm 2 , while the measurements of traditional equipment were mean and median of 0.2 cm 2 and the ICC was 93% (Tables 8 and 9). The results infer that the handheld US is a valuable tool for assisting in the diagnosis of PD as the traditional equipment and therefore a positive response to the study, in keeping with our objective. Toscano et al. (2020) compared the same handheld US used in this study with other non-portable equipment in the area of gynecology and corroborating our result concluded that the handheld US has good accuracy and has potential applicability in routine medical practice.

DISCUSSION
The lowest concordant value of inter-systems measured was that of the midbrain area, it can perhaps be deduced that the midbrain, as it is a sinuous structure, makes it difficult to be measured using a system that TA B L E 7 Transcranial sonography intersystem to assess the quality of TW performed by examiners 1 and 2

Right TW 2 CI 95%
Right  Another limitation is the size of the sample used to assess the inter-system agreement, even so, the study obtained consistent statistical results. The study could not also use only one type of traditional equipment for agreement reference for the handheld US.

CONCLUSION
The handheld US is a viable imaging method for performing TCS because it shows good agreement with the measurements performed with traditional equipment and because its ROC curve and Youden index demonstrate adequate cutoff of the SN area in patients with PD.
The handheld US can be a valuable, practical, fast, bedside method and have the potential to be a cost-effective solution to assist in the diagnosis of PD.