Deciding whether regional nodal irradiation (RNI) works depends upon not only accurate risk assessment but also a better understanding of the disease. Here are some key elements of breast cancer and treatment that aren’t adequately decided just by looking at pathologic stage, grade, and surgery type.
1. Multifocal/Multicentric Disease
Multifocal disease is not part of staging but has a higher risk of LN involvement.1-3 Staging only accounts for the tumors size in a single dimension. Multifocal (multiple areas of disease, same quadrant or section of the breast) and multicentric (multiple areas, multiple quadrants) breast cancer is associated with different biology,1 but it may also have to do with something that standard staging misses—tumor volume. A meticulous study from Memorial Sloan Kettering Cancer Center (MSKCC) showed the risk of lymph node involvement has a stronger link to tumor volume and surface area than it does to T stage (see Figure 1).3
The MSKCC nomogram includes multifocality in predicting additional axillary node metastasis after a positive sentinel node biopsy. Maybe we can predict lymph node metastases better using volume or surface area rather than diameter, or T stage. It is labor intensive and may not be practical, but it may be worthwhile if it becomes easier to assess tumor volume instead of diameter. In borderline cases, it may be worth the extra effort if the risk of nodal involvement will change radiation field design.
Neither NCIC MA.20 nor EORTC 22922 included multifocality as a stratification factor for trial enrollment.4,5 It would be very interesting to see how this variable is linked to outcomes from these trials, as it’s not currently reported in either study.
2. Different Diseases
Molecular research helped identify biologically different types of breast cancer in the early 2000s, recently defined in the St Gallen consensus to guide the role of systemic therapy.6 These subtypes are now commonly defined by combinations of expression of estrogen (ER) and progesterone receptors (PR), HER2/neu, and Ki-67. The risk of lymph node involvement varies based upon these subtypes.7,8 We do know from pre-trastuzumab studies that these subtypes based upon hormone receptors and HER2/neu may have differing local failure rates in breast-preserving treatment.9 Maybe the same is true for lymph nodes.
The EORTC 22922 and NCIC MA.20 trials were designed in 1996 and 1999, respectively, so they couldn’t account for differences in biologic subtypes. Some patients may have received suboptimal chemotherapy, and trastuzumab wasn’t available for many patients. MA.20 allowed trastuzumab but the trial doesn’t report how many participants that affected. The EORTC trial didn’t report it at all.
The MA.20 data show that the 465 ER-negative patients had a 14.6% improvement in 10-year disease-free survival (76.2% vs 61.6%) and trend toward 7.6% improvement in overall survival (81.2% vs 73.9%) with RNI in a group where 91% received chemotherapy. The EORTC trial had only 20% ER-negative participants, versus 40% in MA.20, so it didn’t really report anything on receptor status. However, EORTC showed that patients receiving both hormone therapy and chemotherapy have better survival with RNI (Figure S4 for data geeks). RNI provides a 28% lower risk of dying, which appears to be a crude rate of 4.9% absolute survival benefit for presumably ER-positive patients given both types of systemic treatment.
3. Lymph Node Involvement (Prevalence) Doesn’t Always Mean Recurrence
In the last post on assessing risk of lymph node involvement, I highlighted some numbers that may alarm patients. But the higher prevalence of lymph node involvement doesn’t always mean that the risk of recurrence is just as high.
We’ve seen extremely low rates of regional recurrence in some studies. Some of the older data may be limited because we didn’t have as sensitive ways to monitor for recurrence and some clinical trials only record the first site of recurrence and ignore later ones.
Here is what the NCIC and EORTC trials show for clinical axillary and supraclavicular recurrences:
Both studies only did clinical exam and chest x-rays for routine follow-up, so some recurrences may not have been detected, especially in the internal mammary nodes. Now that we’re doing fewer axillary dissections, maybe the rates will be a little higher even in the highly selected node-positive patients in the ACOSOG Z011 trial (1.5% with SLN, 3.1% with axillary dissection at 5 years).10 But these very small risks of a lymph node recurrence not really explain how RNI can improve survival. I’ll discuss more later.
4. Effectiveness of Chemotherapy
The very low risk of regional failure has been combined with the improving effectiveness of systemic therapy to argue that RNI may not be needed. Improved hormone therapy and chemotherapy regimens have decreased the risk of local-regional recurrences.11 But we also know that clinically silent metastases have a higher chance of recurrence and death.12,13
So how sure are we that chemotherapy really “clears” the lymph nodes enough to leave radiation out, particularly if clinical recurrence rates are low?
NSABP B-18 showed a pathologic complete response (pCR) rate of 31% with AC chemotherapy for clinically node-negative patients.14 A recent European trial, SENTINA, showed that among 360 clinically node-negative, SLN biopsy-positive patients getting chemotherapy, 29% able to undergo repeat SLN and axillary dissection still had residual cancer.15 Patients with residual disease have a worse prognosis,16 and they may be the most likely to benefit from radiation The absence of a clinical nodal recurrence may not be the right measure of success.
5. Biology of Metastasis
Why would there be a possible survival benefit to radiation when the lymph node recurrence risk isn’t that different? In doing a literature search I came up with a provocative hypothesis: because we’re mistakenly evaluating breast cancer the same way for nodes as we do for the breast.
For breast cancer cells to travel to nodes, it is possible that it’s nonrandom which cells migrate, favoring those that preferentially metastasize instead of being locally aggressive.
Multiple studies show nodal metastases appear more aggressive than the primary based upon ER/PR and HER2/neu status.17-20 It may even suggest a worse outcome.19 What if nodal metastatic breast cancer cells aren’t built to grown locally, but instead will spread through the body? That might explain the low regional recurrence rates but still indicate a reason to treat them. We might not see enlarged, firm nodes at all because they now have an evolutionary bias favoring travel over local growth.
Maybe we should be testing nodal metastases for ER/PR and HER2/neu status when we find node-positive disease instead of just retesting the primary tumor at the time of definitive surgery. I haven’t seen breast specialists of any discipline discussing this idea when they debate RNI, but maybe we should. A lot more research is needed to make any conclusions.
6. Effectiveness of Radiation
Radiation therapy won’t do any good if cancer is resistant to it. For BRCA1/2 mutation carriers, radiation appears to be safe and effective.21,22 Looking at the biologic subtypes described above, we know some have higher risks of local recurrence. But cancer subtype radiosensitivity is poorly understood and requires more research to understand how radiation works for each subtype at a molecular level.23,24
We need more research on who, why, and how RNI helps. What should we be measuring for effectiveness? The next post will discuss some of the issues around meaningful endpoints and the importance of how we report the results of research.
This post was originally published for ASCO Connection.
Related posts: It’s Complicated , Assessing Risk , Clinical Trial Reporting , 5 Steps to RNI