An updated systematic review of esthetic grading tools in postmastectomy breast reconstruction

Abstract

Background and Objectives:

We ascertained whether a validated esthetic grading tool for breast reconstruction had been developed and widely adopted since the last published systematic review on the topic from 2015.

Methods:

We performed a systematic review identifying all studies using a grading tool to assess breast reconstruction, using search terms associated with all types of breast surgery and outcomes research. Articles were assessed for patient number, validated scale use, assessor type and training, assessor blinding, assessment method, scoring system type, type and timing of reconstruction, and usage of corroborating scales.

Results:

Of 2809 articles screened, 148 met the criteria. Only 3 used a validated tool, the Esthetic Items Scale. Most used study-only tools (n = 111) or unvalidated tools (n = 28). The most used unvalidated tool was the Garbay/Lowery 5-subscale rubric. Unanchored Likert scales were the most common subjective tool; two-dimensional images were the most used medium. Surgeons, patients, and nurses were the most common assessors. Twenty percent of studies used corroborating scales.

Conclusions:

In the absence of a validated esthetic grading tool for breast reconstruction, researchers continue to rely on unvalidated scales. The only validated scale available is used infrequently and only validated among physicians. A validated, reliable, simple grading tool with clinical and scholastic relevance is needed.

1 |. INTRODUCTION

The esthetic outcome of breast cancer treatment significantly impacts a patient’s well-being and recovery,^1–3 yet there is no single, validated tool to reliably assess the esthetic outcomes of breast reconstruction after mastectomy used in routine practice or research.^4–6 Although several attempts have been made to create such a tool, none meet all five criteria of an ideal grading tool^4,5,7–9: high validity, high reliability, high responsiveness, high interpretability, and low assessor burden.

Two systematic reviews have reported the lack of acceptable, validated tools to assess esthetic outcomes in breast reconstruction patients: Potter et al.⁴ in 2011 and Maass et al.⁵ in 2015. The latter review identified the 10-point scale from the Visser et al.¹⁰ rubric as the highest quality scale available but noted significant limitations and a wide range of intra- and interrater reliability.⁵ Both prior reviews highlighted the lack of consensus as to which esthetic aspects should be considered, or if any should be weighted more heavily than others.^4,5 Furthermore, although the Maass et al. review was published in 2015, it only included articles published until 2013. Nearly 10 years since the last known assessment of the topic, it remains unknown if there is a widely used, validated tool for grading esthetic outcomes of postmastectomy breast reconstruction.

Given the absence of a validated tool, clinicians and researchers have continued to use unvalidated scales such as the Harvard/Harris scale and the Garbay/Lowery scales to evaluate esthetic outcomes of postmastectomy breast reconstruction. More recently, some researchers have substituted patient satisfaction with outcome, a subscale of the BREAST-Q, as a proxy for esthetic outcome.^5,6,11,12 Using unvalidated scales is problematic since many of these scales have not been tested for reliability, validity, and responsiveness; and those that have undergone greater scrutiny are known to have poor reliability, validity, and responsiveness.^5,6 While patient satisfaction is an important clinical metric, it may not correlate with pure esthetic outcome and may be influenced by factors unrelated to the esthetic outcome.^8,13,14 Although BREAST-Q is the gold standard patient-reported outcomes measures (PROM) used in postmastectomy breast reconstruction, how satisfaction on the BREAST-Q correlates with esthetic outcome remains unclear.¹⁴ Patients and surgeons can have different esthetic preferences, which can result in divergent or misaligned esthetic goals.^13,15–18 A validated tool would standardize discussions of esthetic outcomes of breast reconstruction among clinicians^4,5 and would serve as a useful educational adjunct when counseling patients regarding informed decision-making and managing patient expectations.

The purpose of this systematic review of the literature is to examine the use of esthetic grading scales in postmastectomy breast reconstruction, with particular attention to the development and validity of available tools. Our review will serve as an update to the 2011 and 2015 reviews and will critically examine the tools currently in use, setting the foundation for future work in this area of postmastectomy reconstruction care.

2 |. METHODS

2.1 |. Search strategy

We performed a systematic review of PubMed, EMBASE, Cochrane, and CINAHL for all studies before March 2020. Covidence was our primary screening and data extraction tool. Search terms included all controlled vocabulary and keywords related to oncologic, esthetic, and reconstructive breast surgery (including mammaplasty, breast implants, mastectomy, lumpectomy, breast conservation surgery, breast reconstruction, breast reduction, and breast augmentation), and outcomes research (including questionnaires, scales, and surveys reporting patient satisfaction or dissatisfaction, body image, esthetic outcomes, quality of life, self-perception, and appearance). Duplicates were excluded in Covidence. In accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement guidelines, titles and abstracts were screened by two independent reviewers (L. K. and S. W.) per the criteria below. Conflicts were resolved by a third independent reviewer (C. S.). References of the selected articles were manually reviewed to identify additional relevant papers.

2.2 |. Selection criteria

We included full-text English-language articles that described an evaluation of esthetic or cosmetic result in women with postmastectomy breast reconstruction. Articles describing gender-affirmation surgery, other breast surgery (including breast-conserving surgery with or without reconstruction, general esthetic breast surgery, reconstruction for nononcologic purposes, and secondary procedures for improving the appearance of a reconstruction), patient satisfaction without also including an esthetic/cosmetic assessment, or a single aspect of cosmesis (e.g., only nipple-areola complex appearance) were excluded. Abstracts, posters, expert opinions, letters to the editor, and conference reports were also excluded.

2.3 |. Data extraction

Two reviewers performed full-text assessments for data extraction (L. K. and E. P.). Articles were examined for number of patients, whether a validated scale was used or not, who performed the assessment and level of training, whether the assessor was blinded, method of assessment, type of scoring system, type and timing of reconstruction, whether other corroborating scales were used, and whether they met the criteria for an ideal grading tool, as listed previously.

3 |. RESULTS

A total of 4333 studies were identified in the database search, and 1524 duplicates were removed. The remaining 2809 articles were screened, and 965 full-text articles were assessed for eligibility. Of these, 146 studies were included. The other 819 studies were excluded for wrong intervention (n = 436), wrong outcomes (n = 116), abstracts or posters (n = 102), wrong patient population (n = 50), wrong study design (n = 27), ongoing clinical trial (n = 25), language other than English (n = 25), wrong indication (n = 22), redundant study (n = 7), response letter (n = 3), protocol outline or proposal (n = 3), supplement (n = 2), and retraction (n = 1). After hand-searching the references of the full-text articles, 2 articles not identified in the original database search were also included, for a total of 148 articles (Figure 1).

FIGURE 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram of article selection process

3.1 |. Types of scales

Of 148 studies reviewed, only 3^8,19,20 used a validated scale: the Esthetic Items Scale (AIS), which was adapted from the rubric created by Visser et al.¹⁰ (also captured by our review but categorized as study-only scale, as it is not identical to the Dikmans scale) and validated by Dikmans et al.⁸ Of the remaining studies, 111 used a study-only scale/grading rubric, 28 used a named but unvalidated scale, 1 study was a validation of a study-only scale, 3 did not include enough information to conclude whether the scale was validated or not, and 2 used only objective grading criteria (Figure 2). Of the named, unvalidated scales, the Garbay/Lowery 5-subscale method was referenced most often (n = 17), followed by the Harvard/Harris 4-point global scale (n = 11). The other scales referenced include the Sawai/Japanese Breast Cancer Society rubric (n = 2), and scales used by Cohen et al. (n = 2), Salgarello et al. (n = 1), Netscher et al. (n = 1), Brandberg et al. (n = 1), Christie et al. (n = 1), Malata et al. (n = 1), and Tsouskas et al. (n = 1) (Figure 3). Several studies used multiple scales. Subjective grading tools, such as rubrics that did not provide any anchors and did not rely on measurable data, were used as the only means of esthetic assessment in 132 studies (Supporting Information: Figure 1). Two studies used only objective grading criteria, that is, criteria based on measured data (including geometric measurements, applanation tonometry, volume, etc.).^21,22 Thirteen studies used both subjective and objective evaluation criteria. One study did not provide enough information to determine the subjectivity or objectivity of the grading tool used. Objective grading tools included the BCCT.core software (n = 2), the breast retraction assessment (n = 2), applanation tonometry (n = 3), and geometric measurements (n = 10).

FIGURE 2.

Number of unvalidated and validated scales in use

FIGURE 3.

Breakdown of named scales in use

The most common types of subjective tools used were Likert scales (n = 111 studies), followed by numeric rating scales (n = 26) and visual analog scales (n = 11) (Figure 4). Eight studies did not include enough information to determine the type of subjective tool used.

FIGURE 4.

Types of subjective tools used

3.2 |. Assessors

Surgeons were the most common assessors (n = 117), either as sole assessor (n = 41) or as part of a multidisciplinary panel (n = 76), followed by patients (either self-assessments or blinded) (n = 52), nurses (n = 27), residents/fellows (n = 23), students (n = 16), lay people (n = 15), other healthcare providers (e.g., physical therapists, operating room staff, etc.) (n = 13), and non-surgeon physicians (n = 5) (Table 1). Seven studies did not provide enough information to determine who assessed the patients. Eighty-eight studies blinded the assessors to the patients’ identities, 47 studies did not, 11 studies did not provide enough information to determine if the assessors were blinded or not, and 2 studies used both a blinded panel and patient self-assessments (Table 1).

Table 1.

Evaluator training and blinding

Level of training	n
Surgeon	117
Patient	52
Nurse	27
Resident/fellow	23
Student	16
Layperson	15
Other healthcare	13
Other physician	5
Unclear	7
Blinding
Yes	88
No	47
Blinded panel and patient self-assessment	2
Unclear	11

3.3 |. Assessment tools

Two-dimensional photographs were the most common medium used (n = 114), either as the only assessment tool (n = 97) or alongside other media (n = 17). Nineteen studies relied on clinical exams, 13 used patient self-assessments, 5 used dedicated imaging software, 3 used three-dimensional photographs, 1 used standardized patient videos. Eleven did not provide enough information to determine what assessment media were used (Figure 5).

FIGURE 5.

Types of assessment tools used

3.4 |. Reconstruction types and timing

Fifty-seven studies assessed implant-only reconstructions, 50 assessed only autologous reconstructions, and 39 assessed both implant-based and autologous reconstructions (Supporting Information: Table 1). Two studies did not provide information on type of reconstruction. Eighty-five studies assessed only immediate reconstructions, 37 assessed both immediate and delayed reconstructions, 12 assessed only delayed reconstructions, and 9 assessed only 2-stage implant reconstructions (Supporting Information: Table 1). One study specifically assessed only immediate implant reconstruction and 2-stage implant reconstruction, and 1 study assessed only delayed-immediate reconstruction. Four studies did not provide any data on reconstruction timing.

3.5 |. Other corroborating scales

Fifty studies also reported the results of additional tools; 37 of them used PROMs (Figure 6). The most commonly used corroborating tool was the BREAST-Q (n = 18), followed by the Baker capsular contracture scale (n = 8), the short form-36 (n = 6), the European organization for research and treatment of cancer breast cancer-specific quality of life questionnaire (n = 5) and core quality of life questionnaire (n = 3), the functional assessment of cancer therapy-breast (n = 3), the body image scale (n = 3), the Harvard/Harris scale (n = 2), and the Michigan Breast reconstruction outcomes study (n = 2). The disabilities of the arm shoulder and hand questionnaire, the Rosenberg self-esteem scale, the EuroQol-5 dimension, Nahabedian and Tsangaris nipple-areolar complex grading scale, the Gui et al. scale,²³ and the breast cancer treatment outcomes scale²⁴ were each referenced once.

FIGURE 6.

Corroborating scales used. Red bars indicate patient-reported outcome measures (PROMs), blue bars indicate scales that are not PROMs. BCTOS, breast cancer treatment outcomes scale; CC, capsular contracture; DASH, disabilities of the arm, shoulder, and hand; EORTC-QLQ-BR-23, European Organization for Research and Treatment of Cancer breast cancer-specific quality of life questionnaire; EORTC-QLQ-C30, European Organization for Research and Treatment of Cancer core quality of life questionnaire; EuroQol-5D, EuroQol 5-dimension; FACT-B, functional assessment of cancer therapy-breast; MBROS, Michigan breast reconstruction outcomes study; NAC, nipple-areolar complex.

3.6 |. Assessment of grading tools

We also assessed each grading tool based on our previously stated criteria (valid, reliable, responsive, interpretable, low assessor burden). If there was published evidence of validity, reliability, responsiveness, or interpretability testing for a given tool in breast reconstruction patients, and the given tool was deemed to have high validity, reliability, responsiveness, or interpretability score(s) according to the method used to calculate said value, 1 point was given. To gain full points for reliability, a study needed to have both high intra- and interrater reliability; if only one or the other was of an acceptable value, 0.5 points were given. Single scale or global scoring tools were given 1 point for low assessor burden; if a tool had multiple subscales and parameters to consider, 0 points were given.

The highest scoring tools were those used by Dorfman et al.²⁵ with 1.5 points each (both scored 1 point for low assessor burden and 0.5 for reliability testing). Tools in 40 studies scored 1 point, all but one for low assessor burden; the last tool by Liu et al.²⁶ scored 1 point for reliability. Five tools scored 0.5 points for having either acceptable interrater reliability or intrarater reliability, and 102 tools scored 0 points.

No tools scored any points for acceptable validity, responsiveness, or interpretability testing. Forty tools were deemed to have a low assessor burden. The only tool to score 1 point for reliability was by Liu et al.,²⁶ as previously mentioned. Eight studies had either adequate interrater reliability or intrarater reliability; the remaining 139 did not score any points for adequate reliability.

4 |. DISCUSSION

The result of our systematic review concurs with those of the 2011 and 2015 systematic reviews^4,5 that reported the lack of a validated, reliable esthetic grading tool for assessing the esthetic results of breast reconstruction, one of the most critical outcomes following mastectomy, and underscores the need for such a tool. Lowery et al.⁶ initially reported that subscales were more reliable than global scales or visual analog scales, but Maass et al.⁵ found that a 10-point global scale was the highest quality grading tool, in part due to a lower burden on the assessors.

There are five essential criteria to consider in an ideal grading tool: high validity, high reliability, high responsiveness, high interpretability, and low assessor burden. A valid tool measures only esthetic outcomes (i.e., what the tool is designed to measure), not patient satisfaction or only one aspect of cosmesis. A reliable tool produces consistent results over time from the same observers (high intrarater reliability) and consistent results between different observers (high interrater reliability). Responsiveness and interpretability mean the tool is able to distinguish clinically meaningful changes over time and a given score or change in scores has meaningful clinical implications. Low assessor burden means the tool is not overly complicated to administer or understand, and not so time-consuming that it induces survey fatigue.

Although attempts have been made to develop and validate grading tools, including the Cohen et al.⁷ 7-subscale tool and the AIS (based on the Visser et al.¹⁰ 5-subscale tool),⁸ these scales have several shortcomings that limit their clinical usefulness and subsequent widespread adoption. Three studies in our review used the AIS, a validated grading tool to assess esthetic outcomes, which was developed by Dikmans et al.⁸ based on a rubric created by Visser et al.¹⁰ The original Visser scale included five 5-point Likert subscales (volume, shape, symmetry, scarring, nipple-areola complex) and an overall satisfaction score on a 10-point scale.¹⁰ Dikmans added total esthetic score (sum of all subscale scores), and administered the test to plastic surgeons, mammography nurses, and patients for validation. Overall, the AIS was deemed useful for surgeons’ discussions of esthetic outcomes of groups of patients but had limited clinical value for individual patients. Of the three studies that used the AIS, two studies also included PROMs; however, there was low correlation between the AIS and PROM or no correlation could be determined.^19,20

The Cohen scale is a 7-subscale tool that assesses vertical positioning, horizontal positioning, defects, projection, shape, inframammary fold, and overall appearance.⁷ Patients also score their overall satisfaction with their reconstruction experience. Each item is rated on a 5-point Likert scale except for breast positioning, which is scored on a 7-point Likert scale. Cohen attempted to validate this tool in patients and surgeons, but while the test had acceptable to good internal consistency in both groups, the test−retest reliability had a wide range and interrater agreement between patients and surgeons was unacceptable.⁷ The number of subscales also posed a high burden on the assessor, as reported by Maass et al.⁵

Based on this review there is no consensus tool, as demonstrated by the large number of studies that used study-only rubrics. The unvalidated Harvard/Harris and Garbay/Lowery grading tools remain popular, although per our assessment, both scales only score 1 point each for low assessor burden. Furthermore, the Harvard/Harris scale has been used to validate adoption of the BCCT.core—a computer program initially developed for assessing esthetic outcomes in breast conservation surgery and not in postmastectomy reconstruction patients—despite being unvalidated itself.²⁷

The chief weakness of the remaining grading scales we reviewed is that none of them have been validated in patients with postmastectomy breast reconstruction. Some, like the Netscher tool (a 5-point, 4-subscale tool initially designed to assess outcomes after implant removal),²⁸ the Salgarello grading tool (3-point, 7-subscale tool),²⁹ and the Brandberg grading tool (a 7-subscale tool rated from positive to negative for evaluating delayed autologous reconstruction), as well as its modified version,^30,31 have never been formally evaluated or validated. Others, such as the Christie, Tsouskas, and Malata scales were evaluated for reliability but found to have unacceptably low inter- and intrarater reliability.^6,32–34

Another weakness inherent in all the grading tools reviewed was that none had visual representations to anchor each score. They all used text-based descriptions or explanations, which may introduce significant bias depending on the assessor’s level of experience or exposure to breast reconstruction. Clinical scale-validation research in other fields has found that adding anchor images may improve clinical scale constancy; an esthetic grading scale anchored with images representing each score might result in higher inter- and intrarater reliability scores.^35–37

A quarter of the studies in our review also included a PROM, suggesting that some researchers feel that existing grading tools lack such a domain and are attempting to amend the shortfall by including these additional outcomes. The 2011 and 2015 systematic reviews on esthetic assessments of breast reconstruction^4,5 also indicated that a patient-reported domain is a necessary aspect of the ideal grading tool. However, given that most patients allow nonaesthetic aspects of reconstruction to influence their scores despite being instructed otherwise, including such a domain may confuse results instead.^8,13,38

There are several weaknesses in this systematic review. Namely, we did not capture how many studies used patient satisfaction as a proxy for esthetic outcome; however, we know that at least 116 studies were rejected for wrong outcomes and most of those were for measuring patient satisfaction instead of esthetic outcome. We did not extract some parameters related to assessment, namely: how many studies used panels versus individuals, how many observers were used in each study, and what kinds of surgeons were doing the evaluation. We also did not track how many studies used a global esthetic score versus multiple subscales, or which specific parameters were measured. Per our exclusion criteria, we did not capture tools that had been developed or validated in patients who did not receive reconstruction after mastectomy, but if those tools were subsequently referenced in our patient population of interest, they were included and reviewed.

5 |. CONCLUSIONS

There is still no validated esthetic grading tool for breast reconstruction patients. The one validated scale available is not widely adopted and is only validated for use among physicians, not patients. Most researchers continue to rely on unvalidated scales. There is a critical need to develop a validated, reliable, simple grading tool for assessing esthetic results in breast reconstruction patients that has both clinical and scholastic relevance, such as a standard global scoring system anchored with patient images.

DATA AVAILABILITY STATEMENT

The author has provided the required Data Availability Statement, and if applicable, included functional and accurate links to said data therein. The data that support the findings of this study are available from the corresponding author upon reasonable request.