Kenneth Olivier

Dr. Schild has provided an excellent overview of the question of adjuvant therapy in a patient with completely resected stage IIIa NSGLG. The following are my thoughts on how to address the question of whether adjuvant RT should be recommended.

I would first review the findings that led to the diagnosis of stage III lung cancer. Notably, the GT was read as having a suspicious subcarinal lymph node. A subsequent PET scan confirmed the findings on CT, including the suspicious sub-carinal lymph node. The treating physicians proceeded to confirm the radiographic findings with a mediastinoscopy. While in 2007 that would have been the gold standard for mediastinal staging, the traditional cervical mediastinoscopy is not the ideal procedure to sample the subcarinal lymph node station (station 7). The use of endosonogra-phy (transesophageal and transbronchial ultrasound combined with fine needle sampling) allows sampling of almost every lymph node station and is s ignificantly better at staging the mediastinum when combined with mediastinoscopy than mediastinoscopy alone [10]. The mediastinoscopy was negative and the treating physicians made the reasonable decision to proceed with lobectomy, the correct surgical procedure for appropriate management of NSCLC in patients [11].

Then I would carefully review the pathology report. In it, I see that the subcarinal node N2 was “completely replaced” by metastatic squamous cell carcinoma. Aside from “completely replaced,” there is no description of extranodal extension in the subcarinal node. This is reassuring, as extranodal extension may portend a worse prognosis independent of the addition of PORT [12]. Interestingly, the pathology note also questions the bronchial stump margin, but the final summary of the pathology report states that the resection was deemed to be complete with negative margins (R0). The next important consideration is the histology itself. This patient has squamous cell carcinoma, which is worth noting for two reasons. Firstly, the patient nicely fits into the randomized trial conducted by the Lung Cancer Study Group (LCSG) looking at PORT in squamous cell carcinoma of the lung [13], which is discussed subsequently. Secondly, the patient does not have adenocarcinoma, so some of the newer targeted therapies (bevacizumab, pemetrexed) are contraindicated, whereas others (gefitinib and the other EGFR inhibitors) are of diminished utility should he experience recurrence (for a comprehensive summary of this topic see Langer et al. [14]). Although it is difficult to salvage a recurrence of NSCLC regardless of histology, knowing the patient will have a narrowed list of options increases my enthusiasm for more aggressive up front therapy.

Next I would carefully review the operative note. Was the resection of the primary complicated in any way? Did the station 7 lymph node look obviously malignant? Was it fixed to the surrounding tissues? I typically call the surgeon involved in the case and ask if he had any concerns about the margins or degree of resection that might not appear in the pathology note. Hearing a surgeon express concern about the resection would strongly influence me to recommend PORT.

Of note is that the postoperative course for this patient was complicated by a pleural effusion that required talc pleurodesis. We can assume that the pleural effusion was negative for malignant cells, but I would have that fact confirmed. The pleural effusion alone would not change my recommendations for postoperative treatment. Pleural effusions are present in up to 60% of patients after cardiac surgery [15].

Now we are ready to ask the main question. What should our patient do next? My opinion is that the next logical step for the patient is adjuvant chemotherapy. Numerous recent phase 3 clinical trials have independently demonstrated an improvement in overall survival at 5 years with the addition of postoperative chemotherapy [4,16-18]. Specifically, the data support a survival benefit for cisplatin-based chemotherapy so it makes the most sense to recommend drug combinations containing this agent. With respect to delivery, chemotherapy should be sequenced first. I would generally not recommend that this patient receives chemotherapy concurrent with RT, as that combination in the adjuvant setting has been tested in a phase 3 randomized trial and found to be more toxic with no significant improvement in overall survival [19]. The concurrent use of chemotherapy and RT would be of potential benefit only in patients with residual disease after surgery (R1, R2 resections) and that is an interpolated application of the unresected stage III NSCLC data [20].

Do the data support the use of radiation for this patient? As detailed by Dr. Schild, the PORT meta-analysis [1] Published in 1998 showed worsened survival for patients treated with PORT but specifically, for early stage and N0/N1 the negative effect of PORT was more profound, whereas for N2 and stage III patients the best that could be said is that there was no significant difference between surgery and observation. When thinking about the PORT meta-analysis, the main question is why survival would be significantly worse with the addition of PORT. What toxicity of the treatment was shortening the life span of treated patients? The most likely culprit is pneumonitis and pulmonary toxicity attributable to the use of hypofractionated regimens and outdated delivery techniques. The editorial that accompanied the publication included a telling comparison of intercurrent death and biological effective dose (BED) [21] Where the two variables appeared to be strongly correlated.

A study included in the PORT meta-analysis but worth considering separately is the LSCG 773 [13] Trial as was mentioned earlier. It has a specific bearing on our discussion here for a couple of reasons. First, this study avoided some of the pitfalls of other studies included in the meta-analysis by using the more traditional fractionation of 50 Gy in 1.8 to 2 Gy fractions. Second, the study was exclusively of squamous cell carcinomas, which parallels nicely with our patient. LSCG 773 was a negative study in that no survival advantage was seen for the use of RT (regardless of nodal stage), but it did show a dramatic improvement in local control with the addition of PORT. In the observation group, 41% of failures were local failures as compared with only 3% of failures in the RT group. The overall rate of recurrence was also less in the patients with positive N2 nodes who received RT compared with observation (fi = .031).

While this study has a special bearing on our case, we have to keep in mind that our patient is not being treated with observation versus RT. Our patient will have received 4 cycles of cisplatin-based chemotherapy prior to the discussion of RT. From the randomized data presented earlier we can conclude that the benefit of RT is primarily related to a reduction in local failure. The benefit of chemotherapy is the improvement in overall survival, but likely there is some effect on local failure as well. Of the major adjuvant chemotherapy trials, the ANITA trial [4] Reported the patterns of failure such that estimates of the effect of chemotherapy on local failure could be assessed. Local failure was 12% in the chemotherapy arm versus 18% in the observation arm (p = .025). Muddying the waters is the fact that PORT was variably used in this trial, but PORT was more frequently used in observation patients than in chemotherapy patients (33% vs. 22%, p = .0002). So, cautiously stated, chemotherapy using vinorelbine and cis-platin reduces local failure by at least 6% as compared with observation.

As noted by Dr. Schild, the ANITA investigators also published a retrospective analysis of the PORT data used in their trial [22]. It is important to remember that these data are still retrospective despite their origination in a phase 3 clinical trial. Nonetheless, for patients with N2 nodes, PORT improved survival whether they were randomized to chemotherapy (MS 47 months vs. 24 months) or observation (MS 23 months vs. 12 months). There are other data that support a possible survival benefit of PORT in patients with N2 disease at the time of surgery. These include a SEER database study [11] of 7000 patients who received PORT for adjuvant treatment of node-positive lung cancer. PORT prolonged survival for patients with N2 involvement (p = .007). Lastly, we have a phase 3 randomized trial of PORT in patients with N2 disease that is currently accruing patients in Europe. The Lung ART trial is accruing slowly, but the results may answer the question of whether PORT improves survival in this group of patients.

Returning to our patient, what can we accurately say? PORT reliably decreases local failure and that alone is probably worth considering in this patient. If he opts for observation, he may have an 18% risk of local failure (ANITA [22]) and PORT would decrease that to lesser than 5% (LGSG [13]). The modern retrospective data are strongly suggestive that updated techniques may improve survival in N2 patients but the phase 3 data are lacking at this point. When the perennial question is asked “What would you do if it were you?,” I would recommend that he receives PORT. We now need to decide on (a) dose and (b) volumes.

My choice for dose would be 50 Gy in 25 fractions. That was the dose used in the LGSG trial discussed earlier and resulted in outstanding local control. I might consider 50.4 Gy in 28 fractions (1.8 Gy/day), but our patient is 56 years old and I would be considering his need to return to work and routine life. I would like a start date that is at least 2 weeks after the completion of chemotherapy (the recommendation of the ANITA trial) but 4 weeks would also be okay to allow a more thorough recovery from chemotherapy.

Perhaps the more important discussion is what volumes we should treat to a dose of 50 Gy. As we discussed earlier, the toxicity that has given PORT a bad name is likely pulmonary and is strongly related to both dose (which we have decided) and volumes treated. The ANITA trial made no recommendations regarding field arrangement so we have no direction from that data. The LGSG trial required full mediastinal RT in addition to treating the bronchial stump. Do we need to treat large elective mediastinal volumes knowing it will increase the risk of pulmonary toxicity?

In patients with unresectable stage III NSGLG, the use of large field RT has been radically curtailed. Whereas we would have treated a field very similar to the LCSG PORT trial in the 1990s, we are mostly treating “involved field” RT, which places only nodal stations positive on PET, by biopsy, or greater than 1 cm within the clinical target volume (CTV). The reason for that shift was the accumulated data that very few patients experience treatment failure in nodal stations that were not included in the CTV of the involved field. There is ample evidence from prospective clinical trials that use involved field RT [23,24] that the isolated untreated elective nodal failure rate is lesser than 6%. The primary reason the isolated elective nodal is not common is the presence of significant local and distant failure. As we are not getting local control in the regions we directly target, there is little benefit in adding nodal stations that are not obviously involved.

Although it is tempting to want to shrink the fields for PORT in unresectable NSCLC, we should do so cautiously. Our patient has had an excellent local therapy already (surgery) and the addition of PORT will make local failure unlikely. We want to create a CTV that includes the areas at highest risk for local recurrence, and ignores areas at lower risk, remembering we have chemotherapy to help us with local/regional failure as weU. Patterns of failure data are always a great place to start when trying to define a CTV. Investigators at Duke University reviewed the patterns of failure of 61 patients who experienced local failure after an R0 resection for early-stage NSCLC [25]. Overall, the most common site of regional failure was the bronchial stump, making it a priority CTV target in all patients treated with PORT. The authors also nicely lay out patterns of failure for aU the different lobes of the lung allowing for individualized CTVs based on location of the primary lung cancer. For patients with right upper lobe lung cancers, the most common sites of failure were station 4R, the bronchial stump, and station 7. Contouring those sites would be the backbone for an individualized PORT CTV for our patient. Very little has been published on the use of smaller fields in the setting of PORT. There is 1 clinical trial using smaUer-than-traditional RT fields for PORT. An Italian randomized trial of PORT in stage I NSCLC used 50.4 Gy in 1.8 Gy fractions and treated only the bronchial stump and ipsilateral hilum for patients after R0 resection [26]. Local failure in the RT arm was only 2.2% versus 23% for the patients who were randomized to observation. Toxicity was limited to 12% grade 1 pulmonary toxicity for PORT patients. No grade 2+ toxicity was seen. Although the entry criteria for that trial were different than our patient, the principle of covering only high-risk sites is applicable. It would also be reasonable to treat larger, more traditional fields for PORT as long as the dose was reasonable (45-60 Gy) and the bronchial stump was included.

Finally we should discuss the image guidance approach for our patient. A daily cone-beam CT initially aligned to the bony anatomy and refined to cover the bronchial stump would be the optimal approach [27]. This approach decreases the required planning target volume (PTV) expansion substantially and PTV expansions of 5 to 7 mm from the CTV can be appropriate and applied in this scenario. An alternative would be to perform daily kilovolt orthogonal x-rays to align bony anatomy. While simpler than cone-beam CT, there is no method to ensure that the bronchial stump is included in the PTV. Traditional, weekly imaging is also reasonable but requires larger expansions for PTV.

So, in summary, after a long and thorough discussion, I recommend adjuvant chemotherapy for the clear and strong survival benefit associated with its use. I would also recommend the consideration of PORT, firstly for improved local control, and secondly for some possible benefit in overall survival. I recommend a conservative dose of 50 Gy, a limited CTV focusing on high-risk sites, and the use of daily image guidance to decrease the required expansions from CTV to PTV.

  • Contact
  • Category: Cancer