Forty-eight patients met the inclusion criteria, and their medical records were analyzed. The baseline characteristics of the patients are summarized in Table 1. Of the 48 eyes, a liquefied cortex was observed in 20 eyes in the surgical videos. Among the other 28 eyes, bulging of the swollen cortex was observed in six. These findings were not observed in 22 patients. The 26 eyes with these observations were classified as EWC and the 22 without were classified as NWC. The clinical characteristics of the two groups are summarized in Table 1. No significant differences were observed between the two groups regarding sex or ocular comorbidities. However, the patients with an NWC were older than those with an EWC (Table 1, p = 0.0228, t-test).

Of the 48 eyes with a WC and fellow eyes, the appropriate LT was measured using the IOLMaster in 22 eyes (45.8%) and 48 eyes (100.0%: p < 0.001, Fisher’s exact test), respectively. For the remaining 26 eyes where a WC could not be measured appropriately, remeasurement using SS-OCT images was required. In the eyes of remeasurement, the anterior surface of the lens was correctly recognized in all eyes. However, the line indicating the posterior surface of the lens was shifted towards the nucleus, and the LT displayed by the IOLMaster was smaller than the true value.

ACDs were appropriately measured using the IOLMaster in all eyes with a WC and in the fellow eyes. ALs were also appropriately measured in all the fellow eyes. However, ALs in eyes with a WC were appropriately measured in only eight eyes (16.7%). In the other eyes with a WC, IOLMaster showed a failed display and we could not determine the fundus edge in the image of SS-OCT. Therefore, the AL data using ultrasound were used for these eyes.

The LT and CT were significantly greater in the EWC eyes than in the NWC eyes or fellow eyes (p < 0.001, ANOVA). The LT and CT in eyes with an NWC were almost the same as those in the fellow eyes. There were no differences in NT between eyes with an EWC, NWC, or fellow eyes (p = 0.6501, ANOVA). The NT ratios were also significantly smaller in EWC eyes than in NWC eyes or fellow eyes (p < 0.001, ANOVA). The difference in LT (ΔLTs) between eyes with an EWC and their fellow eyes was significantly larger than those with an NWC and their fellow eyes (p < 0.001, t-test).

The ACDs were significantly shorter in EWC eyes than in NWC eyes or fellow eyes (p<0.001, ANOVA). The difference in ACD (ΔACD) between eyes with an EWC and their fellow eyes was significantly larger than the difference between those with an NWC and their fellow eyes (p<0.001, t-test).

There was a significant difference in ALs between eyes with an EWC and an NWC and between their fellow eyes (p=0.0095, ANOVA). However, there was no difference in the ALs between eyes with a WC and fellow eyes for either EWC or NWC.

ROC curve analysis evaluated the detection ability of LT, ΔLT, NT ratio, CT, ACD, ΔACD, AL, and age for EWC with the area under the ROC curve (AUC) values. The AUC values of these parameters except for ACD, AL, and age, were > 0.970, suggesting that these parameters were effective discriminators. Among them, the detection ability of LT (an AUC of 0.977) for EWC was inferior to those of ΔLT (an AUC of 0.991) and CT (an AUC of 0.984) but was almost the same as ΔACD (an AUC of 0.978). The cutoff values of LT, ΔLT, CT, NT ratio, and ΔACD were 5.12 mm, 0.68 mm, 2.05 mm, 61.8%, and -0.30 mm, respectively.

The ROC curves of LT, ΔLT, CT, NT ratio, ACD, ΔACD, AL, and age. LT, lens thickness; ΔLT, the difference in LT between the eyes with WC and the fellow eyes; NT, nucleus thickness; NT ratio, the ratio of NT to LT; CT, cortical thickness; ACD, anterior chamber depth; ΔACD, the difference in ACD between the eyes with WC and the fellow eyes; AL, axial length.

In SS-OCT images, intralenticular clefts, homogenous ground-glass appearance, and any of them were recognized more frequently in eyes with EWC (18/26, 69.2%, 19/26, 73.1%, 23/26, 88.5%, respectively) than in those with NWC (2/22, 9.1%, 0/22, 0.0%, 2/26, 9,1%, respectively; p < 0.001, Fisher’s exact test). Image evaluation to detect EWC in the presence of intralenticular clefts, homogenous ground-glass appearance, and any of them showed a sensitivity of 69.2%, 73.1%, 88.5%, and a specificity of 90.9%, 100.0%, 90.9%, respectively.

CCC was achieved in all cases. In one eye with an EWC, part of the CCC unintentionally extended peripherally. Weakness of the zonule was observed in three eyes (two eyes with an EWC and one eye with an NWC), and a capsular tension ring (HOYA) was inserted. Vitreous strand herniation from the zonule during I/A occurred in one eye with an EWC and was cut using scissors. In the present case, the IOL was inserted into the capsular bag because the lens capsule was intact. There were no other serious complications such as posterior capsule rupture. In all cases, the preoperative visual acuity (LogMar = 1.87 ± 0.06) improved after surgery (visual acuity LogMar = 0.13 ± 0.09, p < 0.001, t-test).