It is widely accepted that surgical resection followed by radiotherapy is a rapid and effective method for treating keloids [11, 15, 16]. Many studies confirmed that surgery followed by radiotherapy could effectively reduce the recurrence rate of keloid and improve the quality of life of patients [9, 14, 17, 18]. However, in the past few years, there was no consensus on the dose and fraction of radiotherapy, and the treatment model varies from different institutes.

Studies have shown that postoperative radiation can significantly reduce the risk of keloid recurrence [7, 17, 19, 20]. A study included 124 patients with 250 keloid lesions reported that surgery followed by radiotherapy with 20 Gy in 5 fractions yielded an excellent local control rate of 94.4% [14]. Another study further verified the effectiveness of the radiotherapy regimen in keloid treatment, with a satisfactory local control rate of 84.8% [20]. Rei et al. reported adjusted radiotherapy regimens according to keloid locations achieved local control rate of 90%, such as 18 Gy radiation dose in 3 fractions to anterior chest wall and scapular region, 8 Gy radiation dose in 1 fraction to earlobes and 15 Gy in 2 fractions to other body site [10]. In this study, we reported comparable cumulative local control rates after a median follow-up of 68.1months. The local control rates of 1 year, 3 years and 5 years were 88.5%, 82.5% and 71%, respectively. This result was consistent with published data [14, 21]. Moreover, the correlation between local control and keloid location was established by our study. The keloid located on chest had the highest recurrence rate, while keloid located on ear had lowest recurrence rate. Similar results were also reported by other study [22]. The prescription BED was similar to other studies, and also achieved a satisfactory effect. But for high recurrence areas, such as the chest and abdomen, perineum, limbs and other high-tension areas, the prescribed BED in this study is slightly lower than some other studies, so this should be one of the reasons for the slightly higher recurrence rates in those areas in this study. It is still unclear why the lesion site can impact the recurrence of keloids. This may be explained by two reasons: ①the skin located on chest always has higher tension which is known to be a risk factor for keloid formation [23, 24]; ②the percentage of collagen, which is higher at chest and back, may contribute to keloid formation [25]. Therefore, the dose and fraction of radiotherapy should be delivered according to the keloid site.

In this study, we firstly reported that symptom of pain and or pruritus were independently prognostic factors for keloid recurrence regardless of keloid site. Although the exact reasons are not clear, these symptoms may be attributed to angiogenesis and fibroblast proliferation [26]. Compared keloid without pruritus symptom, keloid with pruritus always have increased number and density of dermal mast cells and their stored granules, which were considered as important factor to stimulate fibroblast activity and collagen formation [27]. Moreover, patients with high concentrations of inflammatory cells infiltration in keloid tissue were more susceptible to have pruritus symptom [28]. The factors mentioned above are closely associated with keloid formation. That may make it easier to understand why patients with pruritus have higher recurrence rate.

According to literature reports, the occurrence rates of radiotherapy related complications vary greatly and are related to multiple factors [29]. In our study, there were no patients with radiation induced dermatitis, bleeding, and poor wound healing. During treatment and follow-up, infection occurred in 2 patients and cutaneous fibroblastoma in 1 patient. The incidence of infection was significantly lower than that reported in the literature with 4.3-8% [9]. At present, there are few reports of cancer caused by radiation therapy for keloids. Biemans et al. reported a case of fibrosarcoma, which may be related to the carcinogenic effects of radiotherapy. In addition, Raghuvanci et al. reported 5 cases of keloid secondary malignancy after radiotherapy, but there is still no definite conclusion [13]. There were no patients developed cutaneous malignant in the radiotherapy field in this study. However, the cutaneous fibroblastoma was potentially associated with radiotherapy.

There are several limitations for this study. Firstly, the retrospective design of this study may result in bias that impact the accuracy of conclusion. Secondly, the single-center design makes it difficult to widely generalize the radiotherapy protocol. Therefore, a prospective, multicenter, randomized controlled trials are needed to confirm the result of this study.