The interphalangeal angle as a novel radiological measurement tool for Morton’s neuroma – a matched case-control study

Patients were selected which had a MN operatively resected in our clinic between 01/01/2016 and 12/31/2019 and met the following criteria: Inclusion criteria for MN patients were a minimum of 6-month history of neuroma symptoms, a clinical diagnosis of MN and confirmation thereof by MRI. The clinical diagnosis of MN was made by experienced senior physicians of our foot and ankle department and a senior musculoskeletal radiologist by MRI. All included MN were operatively resected and histologically examined in the institute of pathology. All resected samples were histologically diagnosed as MN by an experienced pathologist. Each foot with MN was matched by gender, age and the side of the foot with an asymptomatic control without history of metatarsalgia. The controls were patients of our foot and ankle outpatient clinic, who presented themselves with hindfoot pathologies like ankle distortion, ankle impingement or peroneal tendinopathy and received a weightbearing DP X-ray of the foot. The inclusion criteria for the matched control subjects was a negative history of MN or neuroma-like pain in the forefoot. Exclusion criteria for both MN and control groups were any previous surgery of the foot, any proximal nerve entrapment at the level of the ankle, knee, hip, or back, any history of significant trauma of the forefoot area, any difficulty in walking or standing, diabetes or systemic arthritis. Also excluded were patients with diagnosis of MN simultaneously in multiple webspaces in the same foot, as interdependencies between each other cannot be ruled out. In summary, the final sample consisted of 49 feet of 45 patients. Four patients had bilateral MN. There were 24 right feet and 25 left feet affected. The interspace 3/4 on the right side was affected in 19 feet, the interspace 2/3 on the right side was affected in 5 ft. The interspace 3/4 on the left side was affected in 20 feet, the interspace 2/3 on the left side was affected in 5 feet. The average age of patients with MN and control subjects was 50.7 (range 17.3–73.5) years and 50.6 (range 20.0–73.2) years, respectively. Participant consent was obtained and the study was approved by the cantonal ethics committee of Zurich, Switzerland (BASEC no.2019–01983). The patient characteristics are summarized in Table 1.

Weightbearing DP X-rays of the foot of patients attending the foot and ankle department of our clinic were used in this study. The radiographs were performed preoperatively according to the same scheme of our radiological department. Two assessors performed the radiographic measurements for all patients and controls to check the interrater reliability. Intrarater reliability was assessed on radiographs from 10 randomly selected patients. These were reassessed 4 months after the initial measurements by one assessor. The radiographic measurements were performed via the orthopedic planning software mediCAD (mediCAD Hectec GmbH, Altdorf, Germany). After scaling every X-ray with a 25 mm planning ball, the center line of the diaphysis in the corresponding bone was measured with the function midline through 4 points: Two crosslines (with 2 points each) were made, one in the distal and the other one in the proximal diametaphysal junction in the proximal phalanx. Through these crosslines the measuring tool calculated a midline through the diaphysis. After that, the angle between the midlines of the diaphysis of the affected proximal phalanges was measured to assess the divergence of the proximal phalanges and defined as IPA. Additionally the IMA was measured in the same way (Fig. 1 Radiological measurements). The IMA 1/5, 2/3, 2/4, 2/5, 3/4 and the IPA 2/3, 3/4 were measured. If the lines diverged distally the value was positive, and -vice versa- negative, if the lines converged. The height and width of the MN were measured in the MRI in the coronal sectional view where it had its greatest extension.

The significance level was set at 0.05 and the results are reported as medians and range, unless stated otherwise. A paired t-test sample size estimation yielded a group size of 42 feet (alpha 0.05, power 0.8, minimally clinically important difference and standard deviation from previous research [24]). Case-control matching was performed according to side (exact matching), gender (exact matching) and age (median 0.8 years, maximum 2.7 years). Due to non-normal distributions of the interspace angles, non-parametric tests were used (Wilcoxon-signed-rank test). Inter- and intrarater reliability were measured by intraclass correlation coefficients (ICCs). Due to low numbers of affected feet, no separate receiver operating characteristic (ROC) analysis was performed for IPA 2/3 and IMA 2/3. For the remaining angles, the ROC curve, the Area under the curve (AUC) and its 95% confidence interval were calculated. The AUC was tested by a two-sided binomial z-test. The optimal cut-off value was determined by the minimum distance from the left-upper corner of the unit square. For multiple cut-off values, the diagnostic odds ratio was used as a second condition. This resulted in a unique cut-off value for every index based on the cut-off values, sensitivity, specificity and accuracy. P-Values were not adjusted for multiplicity with regard for sufficient power. The diagnostic odds-ratio was calculated. Association between MN coronal expansion (width and height) in the MRI and angles was assessed by Spearman’s rank correlation. Statistical analyses were computed using Stata/MP 15.1 software (StataCorp LP, College Station, TX, USA).

The diagnostic value of the IPA 3/4 yielded an AUC of 0.75 (p < 0.001), with a sensitivity of 73% and a specificity of 67% and the IMA 3/4 affected by MN compared to controls the AUC was 0.64 (p < 0.05), with a sensitivity of 71% and a specificity of 43%. The IMAs 2/4, 2/5 and 1/5 had no significant diagnostic value (Table 3). A separate calculation of the diagnostic performance for IPA 2/3 and IMA 2/3 was not feasible due to a too small sample size.

The MN had a width of 1.0 mm (0.0; 1.0 mm) and a height of 6.4 mm (3.4; 11.5 mm) in the MRI’s coronal plane. The correlation between IPA (2/3 and 3/4) and MN width and height was non-significant and low at 0.15 (p = 0.316) and 0.08 (p = 0.574), respectively. This was also true for the correlations between the affected IMA and MN width and height which were 0.18 (p = 0.221) and 0.20 (p = 0.161). We found no significant correlation between MN size and IMA 2/4, 2/5 and 1/5.

Several limitations should be considered when interpreting our findings. The results of this study apply to patients who were scheduled for MN excision, which could have favored a selection bias towards bigger MN. Not taken into account were ligament insufficiencies of the MTP joint which might be also a reason for an increased IPA. It should be mentioned that the diagnostic performance of the significantly enlarged IMA 3/4 is moderate. Four patients had a bilateral MN and it is not taken statistically in account, that the right and left foot are highly correlated within the same person. Another statistic limitation is a joint probability of a Type I error higher than 0.05 due to the lack of multiplicity correction.