Keloids are benign proliferative tumors characterized by an excessive cellular response to skin trauma, extending beyond the borders of the original skin lesion. Despite numerous treatment options, there is no single universally accepted therapeutic modality and multiple approaches are often currently employed to improve outcomes and minimize recurrence [1].
Common therapies include: conservative topical management (i.e., silicone coatings, steroids, imiquimod), intralesional injections (corticosteroids, 5.FU, interferon), hyperbaric oxygen therapy, autologous fat grafting, laser therapy, cryotherapy, surgical resection , and radiotherapy [2-10].
Currently, the most widely used treatments for keloids are surgical resection, steroid injection, or radiotherapy [11,12]. Clinical recurrence represents a common indicator of treatment failure.
Surgical resection as monotherapy leads to recurrence rates of 45%-100%, making it an imperfect first-line modality [13,14]. While the use of nonsurgical modalities as adjunctive treatment after surgical resection may lead to a reduction in recurrence, long-term use requirements may make compliance challenging.
Recently, postoperative radiotherapy, as an adjunct to resection, has demonstrated positive early success rates and durability in numerous case series and meta-analyses [1,15,16].
Despite this, current studies may lack adequate sample size, standardized dosing, and long-term follow-up. Additionally, specific characteristics of the keloid, surgical technique, radiation dose, and radiotherapeutic technique may all play a role in keloid recurrence [17].
This study aimed to further examine the effectiveness of medical therapy, surgical resection, and post-resection radiotherapy, in the treatment of patients with keloids, in a high-volume academic tertiary medical center, with more than a decade from experience.
| Methods |
> Study design and population
This retrospective cohort study was approved by the Institutional Review Board of the University of Pennsylvania. The patient population was identified using PennSeek [18], an institutional web tool for searching electronic medical records (EMR). Through PennSeek , adult patients (> 17 years) with the code corresponding to keloid (L73.0) of the International Classification of Diseases, 10th edition, Clinical Modification (ICD-10-CM) were selected , who were treated medically and /or surgically in the Division of Plastic Surgery, or in the Department of Dermatology, within the University of Pennsylvania Health System , between January 1, 2008 and August 1, 2017 (115 months). Those with less than 1 year of follow-up, and patients treated with other topical and oral agents (in the medical management group), were excluded.
> Data collection
Once the patients were identified, the EMR was reviewed by 2 authors and relevant data were obtained from the medical record. To ensure quality control and data integrity, 25 randomly selected medical records were reviewed by 2 additional authors to confirm accuracy.
Then, patients were divided into 3 groups based on keloid treatment: medical management only (MM), surgical resection with or without medical management (RQ), and surgical resection combined with postoperative radiation therapy (RQ+TR).
> Patient characteristics: demographic, clinical and treatment-specific
Covariates extracted from the EMR included demographic data (age, gender, race, ethnicity), clinical factors (smoking status, location, size, number of keloids, recurrence, previous keloid), operative characteristics such as surgical closure (primary, transfer of adjacent tissue, skin graft, others), duration of surgery, and adjunctive use of corticosteroid therapy (pre, intra, or postoperative), and characteristics of radiation, including fraction sizes, number of fractions, biologically effective dose (BED). , a/b = 2.08) [17], and postoperative time until radiation.
> Treatment groups and outcome measurement
MM was defined as keloid treatment limited to topical and/or intralesional corticosteroids (but without surgery or radiotherapy). RQ was defined as standard wide local resection of the keloid, with or without prior or concomitant corticosteroid therapy, but without radiotherapy.
RQ+TR was defined as immediate treatment (within 24-48 hours) with radiation therapy after surgical resection of the keloid. Radiotherapy was administered in a fractionated manner based on the preferences of the radiation oncologist. Treatments involved 3 to 4 sessions on consecutive days with fraction sizes of 300-800 cGy.
The primary outcome measurement for the MM group was changes in keloid symptomatology and appearance, as recurrence was not applicable. Both cosmetic and symptomatic changes were assigned based on documentation in the EMR. The primary measurements in the RQ and RQ+TR groups were recurrence and complication rates.
> Statistical analysis
The unit of analysis was at the lesion level, therefore, patients treated with multiple simultaneous keloids were counted as individual keloids. Descriptive statistics were used to characterize the demographic and clinical data of the keloid.
All categorical variables were compared using chi-square or Fisher’s exact tests (for counts ≤ 10). Normally distributed continuous data were compared using t tests , and non-normally distributed continuous data were compared using the Wilcoxon rank-sum test or the Kruskal-Wallis test, as appropriate.
A Kaplan-Meier curve was used to plot recurrence-free survival between the RQ and RQ+TR groups. A cumulative incidence curve was used to illustrate the time at which half of the patients relapsed (p50).
A multivariate stepwise logistic regression model was used to determine factors independently associated with recurrence and complications, while controlling for confounders.
The covariates included in these models were: age, gender, race, smoking status, and location of the keloid (identified a priori based on its clinical utility and literature), in addition to any factor with p < 0.2 within the univariate analyzes between the RQ and RQ+TR groups. T
All statistical analyzes were performed using Stata/MP software, version 15.1 (College Station, TX). Statistical significance was established a priori at p < 0.05.
| Results |
A total of 284 keloid lesions were included in this study over a recruitment period of 9.6 years, with a median follow-up time of 15.4 months (interquartile range [IQR]: 5.5-30.6). .
The median age of the patients was 39.1 years (IQR: 26.1-53.0). Of the total keloids, 194 were in women (68.1%), 194 were African American (68.1%), and 65 were active smokers or patients who had a significant history of tobacco use (22.8%).
When specific keloid characteristics were searched, 136 (47.7%) lesions belonged to patients with a history of previous keloids, and 95 (33.3%) had a keloid that had recurred after prior treatment.
Two hundred nine lesions were in patients with only 1 keloid (73.3%), while 21 (7.4%) were in patients with 3 or more lesions. The median for keloid surface area was 9.4 cm2 (IQR: 3.5-24).
When baseline characteristics were compared between the 3 study groups, the MM population was significantly older in relation to the RQ and RQ+TR cohorts (41.6 versus 33.4 and 36.8 years, respectively; p = 0.03).
Additionally, the RQ+TR group was more likely to have a history of previous keloids, compared to the MM and RQ groups (73.7% vs 22.1% and 47.9%, respectively; p < 0.01 ), having already had a previous failed treatment (recurrence) (61.1% vs 12.6% and 25.5%, respectively; p < 0.01), and having significantly larger keloids (median 13.8 cm2 vs 6 cm2 and 6.1 cm2, respectively; p < 0.01).
Of the 78 keloids for which cosmetic data were available in the MM group, 30 (25.6%) had substantial improvement, 46 (59%) had slight improvement, and 12 (15.4%) did not improve with the treatment.
Fifty-one lesions in the MM cohort had data on changes in symptomatology with treatment. Of them, 17 (33.3%) had substantial improvement, 20 (39.2%) had slight improvement, and 14 (27.4%) had no improvement. An average of 1.7 and 1.5 injections were required to achieve at least some aesthetic and symptomatic improvement, respectively.
For the RQ group, primary closure was the most common type of closure. The mean duration of surgery was 51.5 minutes (± 36.3). Patients in the RQ group received more corticosteroid therapy pre- and postoperatively than those in the RQ+TR group, p = 0.01 and p < 0.01, respectively.
The median number of radiation sessions per patient in the RQ+TR group was 3.0 (IQR: 3-3). DBE data were available in 95 patients, with a median of 51.1 Gy (IQR: 43.7-69.9).
Sixty-five patients received fractionated doses of 500 cGy or less. When compared to the RQ+TR group, the RQ group had significantly more female patients (75.5% vs 56.8%, respectively; p = 0.01), and a higher percentage of pre- and intralesional injections. postoperative ( p = 0.01 and p < 0.01, respectively).
> Recurrences
Recurrence rates between the RQ and RQ+TR groups were 37.2% vs 37.9%, respectively. Additionally, the mean time to recurrence was also similar between the two groups (290.0 days vs 316.5 days; p = 0.41).
The groups with and without recurrence were compared to determine if there were significant differences in terms of demographic and keloid-specific characteristics. Only gender (54.9% vs 72.9%, respectively ( p = 0.01), differed significantly.
The Kaplan-Meier survival and cumulative incidence curves demonstrated recurrences for both groups, RQ and RQ+TR. 50% of all recurrences occurred at the 12-month mark, and about 80% at the 24-month mark.
After determining potentially significant variables, stepwise multivariable logistic regression was performed to determine which factors significantly affected recurrence and complications.
Male gender and postoperative steroid use were associated with increases in recurrences, with odds ratios (OR) of 3.25 (95% confidence interval [CI]: 1.36-7.76; p = 0.01 ) and 9.54 (95% CI: 4.12-22.06; p < 0.01), respectively.
Keloids that recurred during the study, from both the RQ group and the RQ+TR group, were compared with each other for other differences in demographic and keloid- and surgery-specific characteristics, and complication outcomes.
Positive recurrences in the RQ+TR group had significantly more keloids that had failed previous treatment (recurred) (69.4% vs 34.3%, respectively; p < 0.01), a larger mean size ( 17.0 cm2 vs 6.4 cm2, respectively; p = 0.02), as well as more non-primary surgical closures (44.5% vs 20.0%, respectively; p < 0.01), than positive recurrences in the RQ group.
Positive recurrences in the RQ group had greater preoperative steroid use than in the RQ+TR group (40.0% vs 16.7%, respectively; p = 0.03).
To further investigate the factors leading to recurrence, keloids in the RQ and RQ+TR groups were separated into those that recurred and those that did not recur, and were compared within their respective groups.
In the RQ group, keloids that recurred had fewer female patients compared to those that did not recur (62.9% vs 83.1%, respectively; p = 0.03), and more primary closures than those that did not recur ( 80.0% vs 61%, respectively; p < 0.01), and a higher rate of postoperative steroid therapy than those who did not recur (77.2% vs 39%, respectively; p < 0.01).
Keloids that had failed previous treatment were not more likely to recur after resective treatment. In the RQ+TR group, positive recurrences were more likely than keloids that did not recur to have received postoperative corticosteroid therapy (63.9% vs 13.6%, respectively; p < 0.01).
Keloids that had failed previous treatment were not more likely to recur after radiation treatment. Recurrences had shorter time intervals between surgical resection and the start of radiotherapy than keloids that did not recur (25.0 h vs 29.5 h, respectively; p = 0.01).
> Radiant treatment results
Radiation doses greater than 500 cGy had a statistically significant decrease in recurrence ( p < 0.01), with 33 (91.7%) recurrences in patients who received a dose ≤ 500 cGy/fraction, versus 3 (8, 3%) recurrences in patients who received a dose > 500 cGy/fraction. The administration of radiation was also evaluated as DBE, with a mean of this in the group with recurrence of 51.1 Gy (IQR: 42.7-51.1) versus 51.1 Gy (IQR: 51.1-69. 9) in the group without recurrence, representing a statistically significant difference ( p < 0.01).
> Complications
The RQ+TR group had significantly more complications than the RQ group (17.9% vs 6.3%, respectively; p = 0.01). For both groups (RQ and RQ+TR), recurrent and non-recurrent keloids had similar complication rates. Regarding complications, the differences in the type of closure were statistically significant ( p = 0.02).
The time intervals between surgical resection and the start of radiotherapy did not affect the complication rate. When looking at the effect of dosage or DBE on complications, no statistically significant difference was found. No complications required hospitalization or surgical intervention.
The location of keloids on the extremities and large sizes were associated with an increase in complications, with an OR of 19.51 (95% CI: 3.13-121.77; p < 0.01) and 1.0 ( 95% CI: 1.0-1.0); p = 0.02), respectively.
| Discussion |
Keloids have long posed a treatment quagmire for clinicians, with multiple options available, mixed and inconsistent outcome data, and little objective data on the specific characteristics of keloids that portend favorable treatment outcomes.
This study presents findings from a retrospective review of an almost 10-year experience treating primary and recurrent keloids at a high-volume urban tertiary care center.
The primary outcome studied was the recurrence rate after RQ versus immediate RQ+TR. Previous studies have examined the effectiveness of RQ alone in the treatment of keloids, and found recurrence rates of 30% to 92% [19-24]. Those historically poor results led to interest in other modalities and complementary treatments, including radiation therapy.
An early study by Cosman et al. compared keloid recurrence rates after RQ alone or combined with x-ray therapy (800 r in 4 doses, for 4-8 weeks). They found improved outcomes with the addition of radiotherapy (31.9%), compared to without radiotherapy (48%), fueling interest in radiation therapy [20].
More recent studies looking at RQ combined with immediate radiotherapy have reported recurrence rates ranging from 9% to 71.9% [1,16,21,25-28], also suggesting better results versus RQ monotherapy.
The authors of this study found similar recurrence rates between RQ (37.2%) and RQ+TR (37.9%), respectively. Similarly, time to recurrence for both groups was also equivalent, although there was a trend toward longer times in the radiotherapy group. A possible explanation for this unexpected finding is the nonrandomized distribution of keloids within the treatment groups.
The RQ+TR group had disproportionately higher percentages of patients with a history of previous keloids, an active keloid that had already recurred, and much larger keloids; factors believed to convey poorer treatment outcomes [20]. The RQ+TR group also had more men, which was found to be a significant predictor of recurrence in this study.
It is possible that surgeons recognized keloids that were likely to fail surgical monotherapy and referred those patients for adjuvant radiation therapy. As such, the RQ+TR group had more complex keloids and the expected recurrence rate may be higher. Therefore, similar recurrence may support the superiority of surgery with immediate radiation over surgery alone.
Kal et al. demonstrated decreased recurrence as a function of DBE beyond 10 Gy, with recurrence rates < 10% with DBE values above 30 [29]. The results of the present study correlate well with this, with size fractions of 500cGy or greater leading to a significant decrease in the recurrence rate. Likewise, a higher DBE was associated with a decrease in recurrence, with a mean DBE of 54 Gy.
Since keloid formation is thought to be due to the presence of overly active fibroblasts during early wound healing [17], this result seems expected, as administration of more energy would result in fewer active fibroblasts at the site. of resection and, consequently, a lower propensity to recur.
Additionally, differences in the way radiation is administered can affect treatment results. Electron beam radiotherapy has been shown to be effective in the treatment of keloids [1,30], but other modalities of delivering radiotherapy have been used and may provide superior results [1].
The timing and frequency of radiation can also impact the recurrence rate. The most common dosing schedule at the authors’ institution was 3 sessions on consecutive days with 500 cGy/fraction. Other studies have used similar protocols with good results [30].
One of the biggest concerns with immediate postoperative radiation therapy is an increase in complications. Complication rates were significantly higher in the RQ+TR group versus the RQ group. However, increases in dosage were not associated with increased complications. This suggests that the maximum safe dose of radiation should be used to limit recurrence. Likewise, none of the complications that were documented were major, and all were treated nonoperatively.
Additionally, there was an inverse relationship between time from resection to radiation and recurrence, with shorter delays leading to a statistically higher rate of recurrence. To the authors’ knowledge, this finding has not been demonstrated in the literature and may be a random phenomenon. This is contrary to other studies that have found that prolonged delays lead to greater recurrences [25]. The significance of this is not clear.
MM of keloid lesions is often reserved for patients who are not candidates for other treatments, or who want to minimize pain, time, and complication profile. This study also found that patients treated with MM were more likely to be older than those in the other two groups.
The data demonstrated that there is at least some improvement, both cosmetic and symptomatic, with corticosteroid injections. Therefore, MM is appropriate for those unfit or uninterested in more invasive therapies, given the likelihood of improvement with minimal side effects. Additionally, there is probably no additional benefit after 3 injections.
In this study, it was found that male gender is a risk factor for recurrence (OR 3.25) in both the RQ and RQ+TR groups, independent of other factors.
Other literature that has found that female sex is associated with better results from Nd:YAG laser treatment of keloids [31] is consistent with the results of this study [32]. Neither age, race, location of the keloid, smoking status, history of previous keloid formation, or failure to previously treat a keloid were risk factors for recurrence.
In particular, the finding that location on the head and neck is not a risk factor is surprising, given that it has been linked to a higher rate of primary keloid formation.
Regarding complications , only the location on the extremities and the large size of the keloid were risk factors, with ORs of 19.52 and 1.0, respectively. The significance of the location in the extremity may be due to increased tension during closure due to limited soft tissue.
In the RQ group, the use of pre- and intraoperative corticosteroid injections did not affect recurrence or complication rates. Postoperative corticosteroid injections were not associated with an increase in complications, but were associated with an increase in recurrence. Postoperative injections were identified as a risk factor for recurrence with an OR of 9.53.
While it is possible that postoperative steroids are causally related to increased recurrence, this finding may be related to surgeons being more likely to use them on higher-risk keloids, which were independently more likely to recur.
Another possible explanation is the use of postoperative injections when doctors noticed a recurrence early in the postoperative period. Since any corticosteroid administered after surgery, either 1 day or 1 year later, was pooled, this could bias the pool toward recurrence.
The type of surgical closure affected both the recurrence and complication rates. A possible explanation is that the resection for keloids that could be primarily closed was insufficient, leaving much residual keloid tissue.
The available data are insufficient to prove this, but the result may suggest that more aggressive resections are beneficial to limit recurrence. The type of closure was also associated with greater complications in the RQ group. However, with only 6 keloids in that group, the number may be too small to draw meaningful conclusions.
This study demonstrated that while African Americans were the predominant ethnicity, their treatment outcomes were equivalent to those of other ethnicities. Previous literature has extensively documented the correlation between increased pigmentation and keloid tendency [24]. Similarly, previous manuscripts have found no statistically significant correlation between race and keloid treatments [20].
However, the authors of this work are among the first to compare surgical resection alone with surgical resection with postoperative radiation, using contemporary protocols.
They also demonstrated that keloids that had a previous failed resection treatment do not have a greater probability of recurrence than keloids without previous treatment, or greater complications. This finding means that recurrences after previous resection still need to receive additional treatment, and can probably benefit from surgical resection combined with radiotherapy.
Finally, the Kaplan-Meier and cumulative incidence curves suggest minimal recurrences after the 48-month mark. This may help counsel patients as keloids that have not recurred after 4 years are unlikely to do so.
Strengths of this study include nearly 10 years of experience, large sample size, use of three common treatment modalities, and an overall mean follow-up time of 36.5 months. Previous studies have identified that many recurrences are likely to occur within 13 months [25,33].
With a median follow-up period of 15.4 months, this study likely captured the majority of recurrences. Another strength is the restriction to plastic and dermatological surgeons, to maximize the probability of including only true keloids.
The main limitation of this work involves its retrospective design. A likely consequence is that the surgeons may have altered their surgical plan based on whether the patient had received radiation therapy.
Additionally, patients may have had recurrences but did not seek additional treatment, or were treated at another institution, and were not included in the analysis. Finally, as a single care center experience, the results may not be generalizable to all populations.
| Conclusion |
Radiotherapy remains an effective complementary therapy in the treatment of keloids. Although no significant decrease in the recurrence rate was found among the groups treated with surgery compared with surgery plus radiotherapy, this is probably attributed to radiotherapy being offered mainly to higher-risk patients, with a history of difficult-to-manage keloids.
Although an increase in complications was observed in patients undergoing radiotherapy, these mainly consisted of delayed wound healing and were managed non-operatively. Additionally, higher doses of radiation may be considered as they have been shown to reduce recurrence rates without a consequent increase in treatment complications.
Medical treatment can provide at least some improvement, both in aesthetics and symptoms, but should be reserved for those who are not candidates for surgery.
