Several factors may explain the high prevalence (50.8%; n = 150/295) of cysts found in our population. We used a 3-T MRI with a high magnetic field for increased spatial resolution and a slice thickness on the 3D DP spectral attenuated inversion recovery (SPAIR) sequence of around 0.45 mm, allowing better detection and accurate assessment of very small structures in the millimeter range. Our inter-observer agreement was excellent for detection, volume, and major cyst axis measurements. These findings are better than those reported in the literature [10] based on 1.89 T MRI. We attribute these results to the technical improvements brought by 3-T imaging and 3D image reconstructions with MPR mode. As Lowden et al [10] suggest, another possible explanation for our high prevalence is the “Hawthorne effect”, whereby observers’ behavior changes because of participating in a study. An observer may have searched more intensely for cysts or interpreted the results as showing cysts because of their participation in the study. Nevertheless, the high inter-observer agreement reinforces our belief that observer bias is not the main cause of our high prevalence.

Epidemiological data for our population with cysts are in line with the literature [7, 11,12,13] with an average age of 38.7 years (aged 15 to 75 years) and a sex ratio of 0.6 in favor of females. 88% of our patients were right-handed and the distribution of affected hands was equitable between the two sides. The Pearson correlation test found no significant correlation between the hand affected by the cyst and the dominant hand (r = 0.1691, p < 0.05 see Table 3).

As previously mentioned, most cysts studied were small and discrepancies in measurements between observers were noted for large cysts (see Fig. 4). We explain these differences by the complexity of the anatomy of large cysts. Small cysts are generally perfectly spherical or ellipsoidal and their measurements are more easily reproducible between observers. Large cysts are generally multiloculated with septations and their anatomy is more difficult to assess even when reconstructing in 3 planes with the MPR mode. We cannot conclude statistically about the existence of a correlation between volume ganglion and the existence of acute trauma (r = −0.02604; p = 0.75).

Some authors report the need for a pre-existing underlying scapholunate pathology in the pathophysiology of dorsal cyst development [7]. We found an SL lesion in 39 cases (26%), and although we did not assess the existence of associated scapholunate instability, the number of SL lesions found is close to the 30% of scapholunate instability found in certain arthroscopic studies [7]. However, we cannot draw any conclusions on this point.

The origin of pain in most patients with dorsal cysts is controversial, and several hypotheses have been put forward in the literature, including the possibility of irritation of the dorsal sensory branches of the PIN [2, 7, 14]. We therefore measured the minimum distance between the PIN and the cystic wall (See results). A weak negative correlation was found between distance to PIN and dorsal SL pain (r = −0.2415; p < 0.05), a weak positive correlation between Guérini’s classification and dorsal SL pain (r = 0.2466; p < 0.05), and we cannot conclude statistically about the existence of a correlation between volume ganglion and dorsal SL pain (r = 0.03730; p = 0.65). Although Guérini’s classification is out of date in relation to anatomical knowledge and the discovery of the DCSS, it indirectly reflects cystic volume. All confirm that the origin of pain is complex and probably multifactorial [7], irritation of the PIN and cystic volume alone cannot explain the symptoms.

In the absence of gadolinium injection, we were unable to assess the contrast uptake of the cystic walls and the importance of peri-cystic inflammation in the origin of the pain [7].

Given their benign origin and a spontaneous regression rate of around 50% [2], conservative treatment is the treatment of first choice [3, 15, 16]. Simple aspiration of the cyst may be proposed if it persists, although the recurrence rate is around 59% [17]. Needling techniques [2, 18,19,20], infiltration with hyaluronidase, or the use of intra-cyst corticoids [18] have not shown better results than simple infiltration.

Surgical excision by open surgery or arthroscopy may be proposed in the event of recurrence, with rates of recurrence and complications that vary in the literature, estimated at 21% of recurrence and 14% of complications in open surgery and, respectively, 6% of recurrence and 4% of complications in arthroscopic surgery [3].

We would draw attention to the importance of preoperative MRI in these cases. The use of high-field MRI should provide a precise description of the anatomy of the cyst, its relationship with neighboring structures, in particular the DCSS and DIC, which must be preserved during surgery, and above all the pedicle at the origin of the cyst. The latter should be resected at the same time as the cyst [3]. An adaptation of Guérini’s classification, considering current anatomical knowledge incorporating DCSS, SL pedicle recognition, and location of the cyst in relation to the DIC, would probably be beneficial for anatomical assessment prior to any surgical management, which is why we are proposing a new radiological classification (see Fig. 6). Inter-observer agreement when using this classification was substantial. Using this new classification, we distinguished between very small cysts all located exclusively in the DCSS for which we visualized the pedicle (1b) and those for which the pedicle was not visualized (1a) around 70% n = 72/103 and 30% n = 31/103, respectively. This systematic description of whether or not the pedicle has been visualized by the radiologist using this new classification is therefore essential information to be provided to the surgeon prior to any surgical management, the aim being to complement without replacing arthroscopic observation. In addition, particular attention must be paid to the integrity of the SL and the DCSS, whose lesions will modify the surgical management technique with the addition of a capsulo-ligament repair [21].

Classification of mucoid cysts of the dorsal aspect of the wrist at the expense of the DCSS and the SL ligament. a Type 1a: Cyst originating in the DCSS with no visible pedicle in the SL; b type 1b: Cyst originating in the DCSS with visible pedicle in the SL; c type 2: Cyst originating in the DCSS with extension under the DIC; d type 3: Cyst originating in DCSS extending dorsally to RCD; e type 4: Complex cyst originating in DCSS with extension under DIC and dorsally to RCD. DIC, dorsal intercarpal ligament; RCD, radio-carpal ligament; DCSS, dorsal capsular scapholunate septum; SL, scapholunate ligament

Our study has several limitations. We studied a symptomatic population, which could have had a positive influence on the prevalence of cysts in our population. Furthermore, our study was not designed to provide recommendations for therapeutic management and no follow-up of cysts was carried out.