Results

Clinical description
The demographic, clinical and quantitative MRI findings for the patient group are summarised in Table 1, and for the control group in Table 2.
Table 1 Demographic, clinical and magnetic resonance imaging findings in patients with chronic progressive external ophthalmoplegia related to a single mitochondrial DNA deletion
Gender Age (years) Clinical diagnosis Age of onset (years) Clinical characteristics Mean ROEM (degrees) Mean T2 (ms) Mean EOM area (mm2)
F 19 KSS 12 PEO; ptosis; pigmentary retinopathy; proximal muscle weakness; ataxia; SNHL; fatigue; short stature; low BMI; scoliosis 17.3 80.5 12.3
M 28 CPEO 16 PEO; ptosis; fatigue 10.6 70.1 20.5
M 28 CPEO 14 PEO; ptosis 20.9 67.8 20.4
F 28 KSS 12 PEO; ptosis; pigmentary retinopathy; fatigue; dysphagia; facial, truncal, and proximal muscle weakness; migraine; reduced ET; pain; cramps; migraine; SVT 8.8 85.3 12.7
F 29 KSS 6 PEO; ptosis; fatigue; learning difficulties; SNHL; pigmentary retinopathy; DM; hypertension 25.8 69.5 15.2
F 35 CPEO 14 PEO; ptosis; fatigue; low BMI; proximal muscle weakness; pain 3.2 79.9 15.1
F 36 KSS 17 PEO; epilepsy; migraine; pigmentary retinopathy; proximal muscle weakness 3.1 85.7 18.6
M 45 CPEO 24 PEO; ptosis; fungal keratitis (blind in RE) NA 83.2 15.4
M 52 CPEO 32 PEO; ptosis 16.6 78.0 21.7
Mean values are across left and right lateral rectus, inferior rectus, medial rectus, and superior rectus
BMI body mass index, CPEO chronic external progressive ophthalmoplegia, DM diabetes mellitus, EOM extra-ocular muscle, ET exercise tolerance, KSS Kearns-Sayre syndrome, NA not available, RE right eye, ROEM range of eye movement, SNHL sensorineural hearing loss, SVT supraventricular tachycardia
Table 2 Demographic, clinical and magnetic resonance imaging findings in healthy controls
Gender Age (years) Mean ROEM (degrees) Mean T2 (ms) Mean EOM area (mm2)
F 26 51.5 46.9 32.3
M 28 NA 53.6 36.2
F 28 50.7 56.2 34.1
M 29 45.5 48.9 36.5
F 33 44.1 57.3 29.3
F 35 53.2 57.1 34.9
F 38 49.5 58.4 28.9
F 39 50.2 58.8 32.3
M 43 49.5 58.5 37.6
Mean values are across left and right lateral rectus, inferior rectus, medial rectus, and superior rectus
EOM extra-ocular muscle, NA not available, ROEM range of eye movement

Clinical assessment of ocular movements
The overall mean ROEM was reduced by 73 % in the patient group compared with controls (patients 13.3°, controls 49.3°, p < 0.001; Table 3), with a consistent reduction in individual EOM ROEMs (Fig. 1A). There was no overlap between patient and control range of movement values for any muscle. There was no significant quantitative difference in ROEM detected between the right and left orbit in patients or controls, or between the CPEO and KSS subgroups, when muscles were compared both individually and collectively.
Table 3 Mean range of eye movement, extra-ocular muscle cross-sectional area and T2 relaxation time in healthy control and patient groups
Control Patient Difference
Mean s.d. n Max Min Mean s.d. n Max Min Mean (95 % CI) p-value
ROEM (degrees) 49.3 3.0 8 53.2 44.1 13.3 8.2 8 25.8 3.1 −36.0 (−42.6 to −29.3) <0.001
Area (mm2) 33.6 3.1 9 37.6 28.9 19.2 3.9 9 25.0 13.8 −14.4 (−17.9 to −10.8) <0.001
T2 (ms) 55.1 4.4 9 46.9 58.8 77.8 7.0 9 67.8 85.7 −22.7 (−28.4 to −16.9) <0.001
CI confidence interval, max maximum, min minimum, n number, ROEM range of eye movement, s.d. standard deviation
Fig. 1 Box plot of control (blue) and patient (green) values in each extra-ocular muscle: range of motion (A), cross-sectional area (B) and T2 relaxation time (C). The box represents 25th – 75th percentiles, the solid line in the box represents the median value, stems are to maximum and minimum values, o = minor outlier; + = major outlier. IO inferior oblique, IR inferior rectus, LR lateral rectus, MR medial rectus, SO superior oblique, SR superior rectus

Magnetic resonance imaging analysis
On qualitative assessment, the mean atrophy score in patients was 1.96 ± 0.11, significantly higher than in controls (0.65 ± 0.59, p < 0.001). Only two participants (both patients with CPEO) had focal STIR hyperintensity within EOMs, in 2/12 and 4/12 EOMs, respectively. Focal T1w hyperintensity was more common in patients than controls (72/108 muscles in patients, 49/108 muscles in controls, χ2 = 0.002). In patients this more commonly manifested as a streak (63 muscles) than a dot (nine muscles), whereas in the controls the counts were 25 and 24, respectively (Fig. 2A, 2B, 2C, 2D; Table 4).
Fig. 2 MRI of the orbits in the coronal plane displaying a healthy control on the left column (panels A, C, E, G, I) and a patient with chronic progressive external ophthalmoplegia (CPEO) on the right column (panels B, D, F, H, J). Coronal T1-weighted (panels A and B) and corresponding short-tau inversion recovery (STIR, panels C and D) sequences demonstrate T1-weighted hyperintense and STIR-suppressed foci intrinsic to the extra-ocular muscles (EOMs). If the diameter of the focus was approximately equal in all dimensions in the coronal plane it was termed a 'dot’, whilst if unequal it was termed a 'streak’. An example of a ‘dot’ is demonstrated in the right inferior rectus muscle on paired panels A and C, and examples of ‘streaks’ are shown in both lateral recti on paired panels B and D respectively (arrows). Panels E and F demonstrate an example of the regions of interest (ROI) used to calculate the surface area of each of the EOMs within the right orbit in a control volunteer and CPEO patient respectively: superior-rectus levator complex (SR); superior oblique (SO); medial rectus (MR); lateral rectus (LR); and inferior rectus (IR). Central ROI were placed within the same muscles for the T2-relaxometry measurements in order to avoid contamination by the surrounding fat, as demonstrated in panels G and H within the right orbits of a control volunteer and CPEO patient respectively. Panels I and J demonstrate false colour T2-relaxometry maps
Table 4 Frequency of T1-weighted magnetic resonance imaging changes in healthy control and patient extra-ocular muscles
Controls Patients
Muscle None Dot Streak None Dot Streak
SR 16 2 0 7 0 11
IR 4 2 12 8 0 10
SO 0 14 4 4 6 8
MR 8 4 6 0 0 18
LR 13 2 3 3 2 13
IO 18 0 0 14 1 3
Total 59 24 25 36 9 63
Left and right eye combined
IO inferior oblique, IR inferior rectus, LR lateral rectus, MR medial rectus, SO superior oblique, SR superior rectus
The mean EOM cross-sectional area in patients was 43 % lower than for controls (patient 19.2 ± 3.9 mm2, control 33.6 ± 3.1 mm2, p < 0.001; Fig. 2E and F; Table 3). Muscle T2s were greater in the patient group than controls (75.6 ± 7.0 ms, vs. 55.2 ± 4.1 ms respectively, p < 0.001; Fig. 2G, h, J; Table 3). Consistent individual EOM cross-sectional area and T2 differences were observed (Fig. 1B and C). Quantitative differences in mean ROEM, EOM cross-sectional area and T2 were consistently significant when each muscle was assessed individually (all p < 0.01). There was no significant quantitative difference in EOM cross-sectional area and T2 detected between the right and left orbit in either patients or controls, or between the CPEO and KSS subgroups, when muscles were compared both individually and collectively.

Relationships between clinical and magnetic resonance imaging measures
Correlations between clinical and MRI measures were assessed in the patient group (Fig. 3). There was a significant negative correlation between the patients’ overall ROEM and the mean T2 relaxation time of the EOMs (rho = −0.89, p < 0.01). Mean EOM area did not correlate with disease duration or ROEM. Muscles graded as showing marked atrophy had significantly smaller relative cross sectional area (59 % of control value) than those graded as no (98 % control value) or mild (93 % control value) atrophy (p < 0.001). However, the difference in size between mild and no atrophy was not significant.
Fig. 3 Spearman correlations between clinical and quantitative MRI measures in patients with chronic progressive external ophthalmoplegia (A). A single mean value for each parameter was calculated in each patient for the common regions of interest (superior rectus, inferior rectus, medial rectus and lateral rectus). Range of eye movement (ROEM) correlated with T2 relaxation time. The cross-sectional area did not correlate with either ROEM or disease duration. Scatter plots are shown for T2 versus ROEM (B) and area versus ROEM (C)