Results

Patient characteristics

The cohort included 23 patients (7 women, 16 men; median age 48.0 years). Of the patients, 87.0% were right-handed, and 13.0% were left-handed. In those 3 left-handed patients, the dominance of the left hemisphere was confirmed with functional MRI and neuropsychological testing, while the third patient had solely neuropsychological testing. Surgery was performed using an asleep-awake-asleep protocol in 15 patients (65.2%) and under general anaesthesia with IOM in 8 patients (34.8%). In one of those 8 patients an AF involvement was suspected preoperatively, however an awake procedure was not feasible due to severe preoperative language deficits. In all other cases with presumed AF affection, surgery was performed using an asleep–awake–asleep protocol. Tumor diagnoses included oligodendroglioma CNS WHO grade 2 (39.1%) and grade 3 (17.4%), astrocytoma IDH-mutant CNS WHO grade 2 (17.4%) and 3 (21.7%), and CNS WHO grade 4 astrocytoma in 1 patient (4.4%). Gross total resection was achieved in 47.8% of patients and near total resection in 13.0% and subtotal resection in 39.1%.

See Supplementary Table 1 for baseline characteristics.

Distribution of speech and language deficits

For the definition of the linguistic deficit categories see method section 2.3, 2.6. Postoperative correlation analyses were based on 17 patients with complete language assessment data. At the 3-month follow-up, 11 patients had complete language assessment data for correlation analyses, whereas overall follow-up data were available for 19 patients.

Preoperative assessment

7 of 23 patients (30.4%) had a language deficit. No speech, speech planning, or initiation deficits were observed. Communication deficits were present in 1 patient (4.3%).

Postoperative assessment

Language deficits were present in 13 patients (56.5%), corresponding to 10 patients (43.5%) with a new or worsened language deficit. No speech deficits were observed. Speech planning deficits occurred in 3 patients (13.0%), initiation deficits in 2 patients (8.7%), and communication deficits in 1 patient (4.3%).

3-Month follow-up

Language deficits were present in 7 patients (30.4%). 9 of the 10 patients with new or worsened postoperative language deficits had improved by follow-up, while follow-up data were unavailable for 1 patient. No speech deficits were observed. Speech planning deficits persisted in 1 patient (4.3%), initiation deficits in 1 patient (4.3%), and communication deficits remained present in 1 patient (4.3%).

Fig. 2, Fig. 3 show the distribution of the five linguistic deficit categories across the preoperative, postoperative, and follow-up assessments.

Immediate postoperative language deficits and association with FAT and AF integrity

FAT integrity showed a significant correlation with immediate postoperative language outcomes (Kendall τ= 0.36, p = 0.059). Patients with higher FAT disruption exhibited more severe postoperative language deficits (see Fig. 3). AF integrity showed a moderate, non-significant correlation with immediate postoperative deficits (Kendall τ = 0.30, p = 0.088).

Persistent deficits and association with FAT and AF integrity

Persistent deficits at follow-up were significantly associated with AF integrity (Kendall τ = 0.53, p = 0.033). Complete disruption of AF was most strongly linked to lasting language dysfunction (see Fig. 4). FAT integrity showed a weak, non-significant association with persistent deficits (Kendall τ = 0.10, p = 0.371).

Fig. 4: Distribution of postoperative (left) and follow-up (right) language deficits for the different levels of FAT (above) and AF (below) disruption. The x-axis denotes tract integrity (0 = intact, 1 = partially disrupted, 2 = completely disrupted), while the y-axis indicates the (absolute) affected patient number. FAT/AF score stands for the tract disruption, while 0 is no and 2 complete disruptions in the DTI as described in the method section 2.6. A language deficit was defined as aphasia as described in the method section 2.3 while the darkest color stands for the most severe deficit as described in the method section.

Association of deficits with FAT and AF integrity combined

The summed FAT and AF disruption scores showed a moderate, significant correlation with immediate postoperative deficits (Kendall τ = 0.44, p = 0.022) and lasting deficits (Kendall τ = 0.49, p = 0.046) (Supplementary Table 2).