- © 2000 by American Society of Clinical Oncology
Misconceptions With Staging of Soft Tissue Sarcoma
To the Editor:
We read with surprise the article by Fleming et al in the September 1999 issue of the Journal of Clinical Oncology titled “Long-Term Outcome of Patients With American Joint Committee on Cancer Stage IIB Extremity Soft Tissue Sarcoma.”1
The authors clearly state that T1 tumors are less than 5 cm and T2 tumors are ≥5 cm. This is, in fact, not the case. The fifth edition of the American Joint Committee on Cancer Staging Manual2 and the latest UICC TNM Classification of Malignant Tumors3 for soft tissue sarcoma (STS) both specify that ≤5 cm is the dividing point for T1 tumors and that T2 tumors are greater than 5 cm. This may seem like a minor error on the surface, but when you consider the natural tendency for rounding off around 5 cm (or other size brackets), this can have a significant impact.
Therefore, when Fleming et al claim that stage IIB STSs do not seem to have as favorable an event-free outcome as previously reported, this is not necessarily true. They should state that high-grade STSs less than 5 cm had an unexpectedly low event-free outcome in their series. This is because their article has a notably high percentage of superficial lesions compared with most series of extremity STS, and this makes the wider application of their conclusions inappropriate.
The fourth edition of the AJCC staging manual4 similarly had T1 tumors ≤5 cm; however, even standard STS textbooks present confusing data on this issue.5 The major changes to the fifth edition from the fourth edition of the AJCC staging manual included adding the relatively inconsequential variable of depth in relationship to the investing fascia to create two new substages (stage 1B, low-grade, large superficial tumors, and 2C, high-grade, large, superficial tumors) and an allowance for the relatively good prognosis of small, high-grade STSs (which Fleming et al argue against on the grounds of inadequate follow-up). These changes form the basis of discussions in several recent editorials.6,7
It has also been our experience, after analysis of some subtypes of STS, that high-grade tumors less than 5 cm do not necessarily have a better prognosis than larger tumors.8 However, overall size trend is a continuous variable in STS that is equal in importance to grade.9
We believe the article by Fleming et al1 introduces some important concepts, and further work is appropriate in this area. However, the basic misunderstanding of the current staging system of STS is a symptom of its unsuitability for description of many aspects of STS prognosis determination.7
References
Response
In Reply:
We read with interest the comments of Drs Spillane and Thomas regarding our recent report in the Journal of Clinical Oncology.1 Spillane and Thomas correctly point out that T1 tumor size should be defined as 5 cm or less.2 Although less than 5 cm erroneously appears in our article instead of ≤5 cm as the cut-off point for T1 tumors, we did in fact use ≤5 cm as the tumor size criterion for inclusion in our study.
We would not agree with Spillane and Thomas’s interpretation of event-free outcome in our patients. Their statement that our cohort of patients has “an unexpectedly low event-free outcome” presupposes a specific outcome for these patients. As summarized in the Discussion of our article, the literature on long-term outcome for patients with this stage disease is quite limited. The only large study of T1 soft tissue sarcomas reported to date is that of Geer et al,3 who reported somewhat higher disease-free and overall survival rates at a median follow-up of 42 months than we found at a median of 76 months. Certainly, the longer follow-up in our series and the well-described phenomenon of late recurrence of soft tissue sarcoma4 may contribute to the differences between our observations and the outcome expected by Spillane and Thomas.
Spillane and Thomas state that the distribution of superficial versus deep lesions in our site- and stage-specific report is not comparable to “most series of extremity STS.” Unfortunately, they do not provide any references to indicate which reports they have compared our series to and whether or not those reports were stage-specific. Thus, it is difficult to respond precisely to their statement regarding generalization of our conclusions. Given the significant limitations associated with comparison of single-institution retrospective series from different referral-based populations and the relative paucity of series reporting size-, site-, and stage-specific distributions of clinicopathologic factors, the frequency of superficial versus deep lesions in our report seems relatively similar to stage-specific3 and site-specific reports from other groups.5-8 Nonetheless, we would suggest caution in interpretation of the distribution of prognostic factors in referral-based populations such as our own. It is difficult, if not impossible, to predict the anticipated distribution of prognostic factors within a specific cohort of patients accrued in a referral setting.
We also disagree with the assertion that inclusion of a greater fraction of patients with superficial lesions accounts for what Spillane and Thomas perceive as an unexpectedly low event-free outcome. Indeed, given the well-established better prognosis of patients with T1a lesions,5,6,10 one would postulate exactly the opposite, ie, that our cohort of patients should have improved outcome with the inclusion of a larger fraction of patients with T1a lesions.
With regard to the issues raised by Spillane and Thomas regarding the fifth edition of the American Joint Committee on Cancer Staging Manual, we agree that improvements in the current staging system are needed. This point has also been articulately advanced, with supporting data, in a recent editorial by Brennan.9 We would not, however, agree that depth is a “relatively inconsequential variable.” Tumor depth relative to the investing fascia has been demonstrated to have independent prognostic significance in several recent prognostic factor analyses.5,6,10 Spillane and Thomas may be less familiar with this because depth was omitted from a recent analysis of soft tissue sarcoma prognostic factors from their institution.11
We agree with Spillane and Thomas that important observations can be made from retrospective studies. In this regard, the most important observations from our study seem to be: (1) there is a subset of patients with T1 high-grade sarcomas who have a poor prognosis that seems to approach that of other patients with traditional high-risk extremity sarcomas, and (2) long-term follow-up is important (even for patients with ostensibly early-stage sarcoma), since 15% of recurrences were observed beyond 5 years of follow-up.
