The Agreement of the Nomogram Tool and Ultrasound Biomicroscopy Images in Calculating Ultrasound Cycloplasty Probe Model in Chinese Patients

With the increasing aging of China, the prevalence and burden of glaucoma will continue to increase in the next 30 years [1]. Currently, the treatment of glaucoma focuses on controlling the intraocular pressure (IOP). Traditional surgical treatment is represented by filtration surgery, which mainly reduces the IOP by promoting aqueous outflow. Although it has a significant effect on the IOP reduction, it has the disadvantages of high conjunctiva status requirements and postoperative scarring, which leads to IOP recovery [2]. Cyclodestructive surgeries such as cyclophotocoagulation and cyclodiathermy deliver nonselective energy to damage the ciliary to reduce the IOP production and damage adjacent structures, which causes inflammation, severe pain, and even visual loss [3‒5].

The ultrasound cycloplasty (UCP) procedure has been highlighted in recent years due to its merits, including noninvasion, high security, and repeatability. Using high-intensity focused ultrasound, the UCP procedure achieves accurate and controlled localization of energy to the ciliary body with little absorption by other tissues [6]. The safety and effectiveness of UCP have been demonstrated in several clinical studies in Europe, India, and China over the past 10 years [7‒11].

During UCP surgery, ultrasonic energy can be delivered to the ciliary body through probes with three different models. Choosing the right model of the probe for the patient is one of the most important steps to accurately localize ultrasonic energy to the ciliary body. Three probe models are currently available: models 11, 12, and 13. The calculation of probe models was determined by ultrasound biomicroscopy (UBM) images in earlier research and clinical application [6]. With the number of cases undergoing UCP increased, a nomogram tool to predict the model had been structured based on the biological parameters of European patients who had undergone the UCP procedure, the two parameters were axial length (AL) and white to white (WTW) [12]. When using UBM images to calculate the transducer model, engineers need to scale the UBM image with clear ciliary body according to the length of the ruler, compare it with the probe diagram of three models, and compare which model of probe can concentrate more ultrasonic energy on the ciliary body. This method is accurate but time-consuming and demanding. When using the nomogram tool, you only need to enter the patient’s AL and WTW, and you can immediately get the corresponding probe model. Compared with the method of using UBM images, using a nomogram tool to calculate the model is more convenient and faster and does not require clear UBM images with clear ciliary body imaging.

However, there are no studies that have confirmed the accuracy of the nomogram tool to calculate the model of the probe, especially for different races like Chinese glaucoma patients. In this study, we collected a sufficient number of relevant biological parameters from Chinese glaucoma patients suitable for UCP surgery, used the two methods to calculate the probe model, and analyzed their results in calculating the probe model to explore the agreement between the two methods of calculation for Chinese glaucoma patients.

This study recruited glaucoma patients who visited Zhongshan Ophthalmic Center in Guangzhou from January 2019 to December 2019 and were eligible for UCP surgery, including glaucoma patients (1) aged 18–90 years; (2) with IOP higher than 21 mm Hg under medication; (3) failed to control IOP to target levels by conventional glaucoma surgery; (4) diagnosed with primary open angle glaucoma (POAG), primary angle-closure glaucoma, secondary glaucoma (SG), or congenital glaucoma (CG). Exclusion criteria included eye infections 2 weeks before surgery, aphakia, narrow palpebral fissures, hollow eyes, thyroid-associated ophthalmopathy, history of ciliary body surgery, vitrectomy, staphyloma, intraocular tumor or retrobulbar tumor, pregnancy, and any serious systemic diseases. All participants underwent the following baseline examinations: best-corrected visual acuity, IOP measurement, Humphrey visual field test, slit-lamp biomicroscopy, fundoscopy, UBM (model SW-3200L; Tianjin Suowei Electronic Technology Co., Ltd., Tianjin, China), and IOLMaster 500 (Carl Zeiss Meditec, Jena, Germany; software version: version 5.4). This study was approved by the Independent Institute Research Ethics Committee at Zhongshan Ophthalmic Center (ZOC, Guangzhou, China) and adhered to the tenets of the Declaration of Helsinki.

EyeOP1 device developed by Eye Tech Care company was used in the study. UCP treatment is based on therapy probes that generate focused ultrasound energy. According to the diameter of the therapy probe (11-, 12-, and 13-mm), EYEOP-PACK was called models 11, 12, and 13, respectively. When using the UBM image to measure the probe model, the ciliary body must be clearly imaged. By simulating whether the ultrasonic energy of the three different types of probes is positioned on the ciliary body, the appropriate model can be selected. UBM images of all patients were sent to the same trained technical personnel for probe model measurements. The AL and WTW biological parameters measured by IOLMaster 500 (Carl Zeiss Meditec AG) were plugged into the nomogram tool to obtain the probe model.

All data were recorded in Microsoft Excel spreadsheets and analyzed by SPSS for Windows version 28 software (SPSS Inc., Chicago, IL, USA). The visual acuity, IOP, ocular AL, and horizontal corneal diameter (WTW) are expressed as the mean ± standard deviation with the range. Three probe models of these Chinese patients calculated by the UBM image and nomogram tool were cross-tabulated. Agreement between the results was assessed by calculating the unweighted kappa (κ) and weighted (using a linear scheme) kappa (κ) values. According to Landis and Koch’s guideline for the strength of agreement, which is indicated by weight κ values ranging from −1 to +1, κ<0 is considered poor agreement; 0–0.20: slight agreement; 0.21–0.40: fair agreement; 0.41–0.60: moderate agreement; 0.61–0.80: substantial agreement; 0.81–1.00: almost perfect agreement [13].

In total, 1,061 eyes of 642 patients with 5 types of glaucoma, which included POAG, acute primary angle-closure glaucoma (APACG), chronic primary angle-closure glaucoma (CPACG), SG, and CG, were enrolled in this study. Table 1 describes the characteristics, best-corrected visual acuity, IOP, AL, and WTW of all patients. Their mean age was 61.66 ± 11.66 years old, ranging from 18 to 87 years old; 362 (56.39%) females and 280 (43.61%) males were involved. AL and WTW were all measured by well-trained ophthalmologists using the same IOLMaster 500 device. The mean AL was 22.88 ± 1.33 mm, ranging from 19.15 to 32.14 mm. The mean WTW was 11.52 ± 0.49 mm, ranging from 10.00 to 12.90 mm.

Table 2 summarizes the agreement between the two methods to calculate the probe models of Chinese glaucoma patients. The overall rate of exact agreement is 62.86%. The weighted κ is 0.299, and its 95% CI is 0.258–0.340. By accessing the weighted κ, the level of agreement in all patients was fair.

Figure 1 shows the distribution of the three models by using UBM images to calculate and the number of smaller, equal, or larger models by using the nomogram tool in each model, compared with the results by UBM images. The white area showed the number of larger models by the nomogram tool than by UBM images, while the black area showed the number of smaller models by the nomogram tool than by UBM images, and the gray area showed the number of the same model by two methods. Compared with UBM calculations, the nomogram tool preferred to use larger probe models, and 68.65% of the inconsistent results by two methods were due to nomogram tool calculations being larger than UBM calculations. Particularly in the eyes of the UBM image calculated as the 11-probe model, 155 of 177 are calculated as larger models by the nomogram tool.

Patients diagnosed with 5 types of glaucoma are distributed as shown in Figure 2. APACG patients accounted for the most patients (36.14%), followed by CPACG patients, with a proportion of 31.78%. Patients with POAG, SG, and CG accounted for 21.81%, 9.81%, and 0.47%, respectively.

Taking the types of glaucoma as the classification, the probe models by two methods of the 5 types of glaucoma patients were cross-tabulated, and κ statistics are summarized in Table 3. Since there were only 4 eyes in CG patients and the probe models calculated by the two methods were both model 13 for the four eyes, κ statistics were not available. The weight κ values in POAG, APACG, CPACG, and SG patients were 0.336, 0.146, 0.204, and 0.332, respectively.

The aqueous humor is produced by the unpigmented epithelium of the ciliary processes, while the pigment epithelium of the ciliary body is rich in pigment. Thus, precisely changing the state of humor secretion of the unpigmented ciliary epithelial cells of the ciliary process is the key to cyclodestructive surgeries. Unlike optical energy, which can be absorbed by pigment tissue, high-intensity focused ultrasound delivered by the miniaturized transducers enables one to control the energy in a smaller focal zone and unpigmented tissues [14‒16].

The exact agreement is 62.86%, the simple κ is 0.264, and the weighted κ is 0.299. Thus, if a nomogram tool is used to calculate the probe model for glaucoma patients in China, 37.14% of the results may not be accurate enough. According to Landis and Koch’s guideline for strength of agreement indicated with weight κ values, two methods to calculate the probe model of Chinese patients were at a fair agreement level. The undesirable consistency is related to the principle of measurement by both methods. Using UBM images to calculate the probe model is to put it on model diagrams of probe models 11, 12, 13 and check whether the ultrasonic energy delivered by the miniaturized transducers is exactly focused on the ciliary body [6, 16]. Since this method can accurately determine whether the ultrasonic energy focusing region of probes with different models and the ciliary body area match so that it simulates the accurate focusing of the energy of a particular probe on the ciliary body during the UCP surgery, we believe this method has a higher degree of accuracy but it requires high imaging clarity for the ciliary body and requires professional technicians using model diagrams for measurement, which means this method is time-consuming and not so convenient. In order to calculate the model of the probe more conveniently and rapidly, the nomogram tool was structured [12]. It was built by performing multivariate regression analysis on the data of probe models and the AL and WTW of a sufficient number of patients who underwent UCP surgery. In the nomogram tool, according to the values of AL and WTW, three ranges are delineated, corresponding to probes 11, 12, and 13, respectively. When calculating in this way, only the data of AL and WTW need to be entered and the corresponding probe model can be shown immediately, which makes it fast and convenient. But based on WTW, the data in the horizontal direction of the eye may be a factor contributing to the discrepancy with the UBM measurement because the ultrasonic energy of the UCP needs to avoid the three- and six-point orientations and that is why images in the horizontal are usually avoided to be used in the UBM measurement [6]. Another possible factor for the different results in calculation of the two methods may be that the data used to build the nomogram tool comes from a single ethnic source. The relationship between ciliary body spacing and position and AL and WTW may vary by race [17, 18]. So when calculating probe models for glaucoma patients of other races, the nomogram tool may have some deviations. In addition, a third factor that may contribute to the lack of consistency between the two methods may be that both methods have a limitation that they only use data from just a single direction of the eye, but during the UCP procedure the ultrasonic energy is focused in 3, 4, or 5 directions (corresponding to 6, 8, 10 sectors) [19]. This factor may cause errors not only in the nomogram tool but also in the UBM measurement. For the above reasons, the UBM image measurement probe model method is still more recommended. But this method is expected to be improved by calculating UBM images in more directions. And for the nomogram tool, it is expected to include data from more ethnicities to make it more universal.

68.65% of the inconsistent results (265 of 386 as shown in Fig. 1) were due to nomogram tool calculations being larger than UBM calculations, showing that the nomogram tool preferred to use larger probe models, especially in model 11, in which 87.57% (155 of 177) are calculated as larger models by the nomogram tool. The nomogram tool using more larger probe models may affect the effectiveness of the surgery, but this design may be out of consideration of ensuring the safety of the patient’s lens during surgery. Because the energy range of larger probes is more distant from the lens and studies had proved that the lens has a very high acoustic absorption coefficient and thus are more likely to be impaired during the UCP procedure. But at the same time, it can pose a potential threat to the ora serrata part if the eye is not in the right position during the surgery [19, 20].

All patients were classified based on the glaucoma type, and their exact agreement, simple κ value, and weighted κ were calculated. Figure 2 shows the distribution of patients diagnosed with 5 types of glaucoma in this study. Table 3 showed that the 2 methods in POAG patients have the largest weighted κ value, which indicates that the level of agreement is fair. The weighted κ values are only 0.146 in the probe model agreement calculation of APACG patients, which means their agreement is slight. The weighted κ value of APACG patients is the lowest among 5 types of glaucoma patients. And the value of the weighted κ value of CPACG patients is between that in APACG and POAG patients. The difference of the weighted κ value may be related to the anterior rotation of the ciliary body in patients with APACG and CPACG. Studies have shown that compared with CPACG, APACG had more anteriorly positioned ciliary body. Greater degree of ciliary anterior rotation leads to more uncertainty in predicting ciliary position with AL and WTW, while in patients with POAG, there is no significant change in ciliary position and fewer uncertainties in prediction, which may lead to a higher agreement between the two methods [18, 21, 22]. The weighted κ values of SG are very close to those of patients with POAG. In SG, the causes are varied, and causes that could lead to anterior rotation of the ciliary body account for a relatively small proportion [23].

This study was a hospital-based retrospective study, which may cause bias at the baseline. These results may need to be validated in a multi-center, broader group of Chinese glaucoma patients.

In summary, in the present study, a nomogram tool and UBM images of Chinese glaucoma patients suitable for UCP were used to calculate their probe models. By analyzing the calculation results, we found that the agreement between the two methods were at a fair agreement level in Chinese patients with glaucoma. The nomogram tool prefers to use larger probes. The nomogram tool is expected to include data from more ethnic groups and be able to calculate separately for different types of glaucoma. In addition, the inclusion of parameters or images from more directions of the eye can help measure probe models more accurately for both the nomogram tool and the UBM image measurement.

This study was approved by the Independent Institute Research Ethics Committee at Zhongshan Ophthalmic Center (ZOC), approval number (2021KYPJ083). The written informed consent was obtained from all participants in accordance with the Declaration of Helsinki. All experiments were performed in accordance with relevant guidelines and regulations.

The authors have no conflicts of interest to declare.

This study was supported by the National Natural Science Foundation of China (81970808) and the Natural Science Foundation of Guangdong Province (2019A1515011196, 2020A1515010121).

C.Z.: contributed to the study concept. M.L. and C.Z.: contributed to the study design. D.W., L.C., Z.H., Y.T., and Y.L.: collected the patient data and medical records. Z.W. and L.J.: contributed to the data interpretation. S.Z.: wrote the manuscript. All authors contributed to the critical revision of manuscript and the final approval for its submission.

The data used to support the findings of this study are available from the corresponding author upon request. Data are not publicly available due to ethical reasons. Further inquiries can be directed to the corresponding author.