chemoradiotherapy for non-small cell lung cancer correlates with patterns of failure

Lung Cancer (2005) 49, 95—108

Metabolic (FDG—PET) response after radical radiotherapy/chemoradiotherapy for non-small cell lung cancer correlates with patterns of failure夽 Michael P. Mac Manus a, ∗, Rodney J. Hicks b, Jane P. Matthews c, Andrew Wirth a, Danny Rischin b, David L. Ball a a

Department of Radiation Oncology, Peter MacCallum Cancer Centre, St. Andrew’s Place, East Melbourne, Vic. 3002, Australia b Department of Nuclear Medicine, Peter MacCallum Cancer Centre, Melbourne, Vic., Australia c Statistical Centre, Peter MacCallum Cancer Centre, Melbourne, Vic., Australia Received 31 July 2004 ; received in revised form 25 November 2004; accepted 29 November 2004 KEYWORDS Positron emission tomography; Non-small cell lung cancer; Survival; Metastasis; Radiotherapy; Chemotherapy

Summary Background: We previously reported that F-18 fluorodeoxyglucose (FDG) positron emission tomography (PET) response correlated strongly with survival after radical radiotherapy (RT)/chemoradiotherapy for non-small cell lung cancer (NSCLC). PETresponse, survival and patterns of failure data are presented with long-term followup. Methods: Pre- and post-treatment FDG—PET scans were performed for 88 patients after concurrent platinum-based radical chemo/RT (n = 73) or radical RT alone (n = 15). PET responses were prospectively assessed as either complete metabolic response (CMR), partial metabolic response (PMR), stable metabolic disease (SMD), or progressive metabolic disease (PMD). Results: RT was 60Gy in 30 fractions in 6 weeks. Follow-up PET was performed at a median of 70 days after treatment. PET responses were: CMR, n = 40 (45%); PMR, n = 32 (36%); SMD, n = 5 (6%) and PMD 11 (13%). Estimated median survival after follow-up PET was 23 months; median follow-up duration 35 months. One and 2 year survival after follow-up PET was 68% and 45%, respectively. Median survival for CMR and non-CMR patients was 31 and 11 months, respectively (p = 0.0001). One-year survival for CMR and non-CMR patients was 93% and 47%, respectively and 2 years survival was 62% and 30%, respectively. Excluding PMD patients, non-CMR

夽 These data were presented at the 45th Annual Meeting of the American Society for Therapetic Radiology and Oncology, Salt Lake City UT, October 2003. * Corresponding author. Tel.: +61 3 9656 1111; fax: +61 3 9656 1424. E-mail address: [email protected] (M.P.M. Manus).

0169-5002/$ — see front matter © 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.lungcan.2004.11.024

96

M.P. Mac Manus et al. patients had higher rates of local failure (HR 2.15, p = 0.009) and distant metastasis (HR 2.05, p = 0.041) than CMR patients. By last follow-up, 20 of 40 CR patients (50%) had PMD, with local failure (n = 8), distant metastasis (n = 2) or both (n = 10). Conclusions: Attainment of CMR after radical RT/chemoRT for NSCLC bestows superior freedom from local and distant relapse; late local relapse is common. © 2005 Elsevier Ireland Ltd. All rights reserved.

1. Introduction It could be invaluable to have an early indication of the success of therapy in patients with non-small cell lung cancer (NSCLC) treated with definitive radical radiotherapy (RT) or chemoRT. If it was possible to anticipate future patterns of disease progression, selected patients could potentially benefit from additional therapy. CT scanning is currently the standard method for response assessment after radical RT but it suffers from well-known limitations. These limitations include dependence on the unreliable parameter of lymph node size to determine involvement by tumor [1] and an inability to distinguish between inactive scarring [2] or necrotic tumor and active tumor after therapy [3]. These limitations are exacerbated by the presence of radiation pneumonitis [4] and atelectasis. In a previous study of 73 patients with NSCLC, we demonstrated that positron emission tomography (PET) using 18 F-fluorodeoxyglucose (FDG) is superior to conventional imaging with CT scanning for response assessment after radical RT/chemoRT [5]. Response assessments made with FDG—PET and CT were identical in only 40% of cases. PET showed a much higher rate of complete response than CT. Additionally, PET response (or metabolic response) was much more powerfully correlated with survival than CT response using WHO criteria [6], stage or Eastern Co-Operative Oncology Group (ECOG) performance status (p < 0.0001). There have to date been no published data on the long-term survival prospects for patients who attain complete or partial metabolic responses to RT, although it is already clear that those in the progressive disease category typically have very short survival. Patients who attain complete metabolic response on PET have no detectable disease and in our earlier report had superior survival compared to those in other response categories. However, because PET cannot detect microscopic volumes of residual disease, prolonged follow up is required to estimate the proportion of complete responders who ultimately attain durable locoregional disease control. Similarly no data have previously been published on the patterns of failure that occur in the different metabolic (PET) response categories.

In this paper, we describe the patterns of failure seen in a cohort of patients who underwent restaging PET scans after radical RT/chemoRT for NSCLC. This series includes the 73 patients described in our initial report and updates their survival. It also includes an additional 15 NSCLC patients treated with radical RT/chemoRT who had both staging and restaging PET scans.

2. Methods 2.1. Patients FDG—PET scanning was performed as part of a prospective study of the value of functional imaging in NSCLC at the Peter MacCallum Cancer Centre (PMCC). This study was initiated in 1996, was approved by the institutional clinical research and ethics committees and was in accordance with the Helsinki declaration of 1975, revised 1983. Patients described in this report had NSCLC treated with radical radiotherapy or chemoradiotherapy and had undergone both staging and elective post-treatment 18 FDG PET scans. All eligible patients with post-treatment scans performed between May 1997 and September 2000 were selected for analysis. There were 88 such patients. In 15 cases, a recent restaging CT scan was unavailable for comparison with the PET scan. However this analysis is confined to the PET results. The revised staging system described by Mountain was employed [7]. In the first year, unconfirmed PET-apparent extensive disease did not preclude an attempt at radical treatment for patients who were otherwise eligible. With confirmation of the high accuracy of PET, it was subsequently considered unethical to give radical therapy to patients with previously unsuspected PET-apparent extensive disease. Follow-up imaging was performed following completion of radical radiotherapy after a planned interval of 4—12 weeks. Scans were occasionally performed outside these limits because of unavailability of FDG or transportation problems for patients residing at distant locations.

Patterns of failure after PET in NSCLS

2.2. Treatment policy Patients with good performance status and without significant weight loss with medically or surgically unresectable stage IA-IIIB NSCLC were offered radical radiotherapy and concurrent platinum-based chemotherapy, provided disease could be treated with a tolerable radiotherapy target volume. Reasons for use of radiotherapy rather than surgery in stage I—II disease included severe chronic obstructive pulmonary disease and significant cardiovascular disease. Patients with ECOG performance status 2 or significant weight loss were considered for radical treatment if other prognostic factors were favorable. Patients refusing chemotherapy or with inadequate renal function received radiotherapy alone. The target volume, comprising the primary tumor, clinically involved mediastinal and hilar nodes and adjacent uninvolved ipsilateral mediastinal nodes, was treated to 60Gy in 30 fractions in 6 weeks.

2.3. PET scanning acquisition and processing A GE Quest 300-H PET scanner was used (UGM Medical Systems Inc., Philadelphia, PA, USA). Emission data were processed using iterative reconstruction, with and without measured attenuationcorrection. Scans were performed with arms raised, in the radiotherapy treatment position, and encompassed the lower neck, thorax and upper abdomen.

2.4. Assessment of treatment response Pre- and post-treatment PET scans were coregistered and examined (a) for response at sites of known disease and (b) for evidence of progression in the thorax or at distant sites, almost always by a single reader. The PET physician was blinded to the results of the follow-up CT. PET-apparent radiationinduced pulmonary and pleural inflammation was noted. Using a standardized display to provide a consistent intensity of background soft tissue activity, PET scans were scored for response using the following criteria: (1) Complete metabolic response (CMR): No abnormal tumor FDG uptake; activity in the tumor absent or similar to mediastinum. (2) Partial metabolic response (PMR): Any appreciable reduction in intensity of tumor FDG uptake or tumor volume. No disease progression at other sites.

97 (3) Stable metabolic disease (SMD): No appreciable change in intensity of tumor FDG uptake or tumor volume: no new sites of disease. (4) Progressive metabolic disease (PMD): Appreciable increase in tumor FDG uptake or volume of known tumor sites and/or evidence of disease progression at other inrathoracic or distant metastatic sites. Radiation-induced inflammatory changes in the lungs and pleura had a different distribution from tumor uptake and were not scored as persistent or progressive disease. These inflammatory changes conformed to the volume of irradiated lung and could be readily distinguished from persistent tumor uptake by their location and pattern of uptake.

2.5. Follow-up after treatment Most patients were followed up by physicians at PMCC or one of its satellite facilities. Patients were seen at 2-monthly intervals for the first 2 years with 2-monthly chest radiographs. After assessment of treatment response, other investigations, including CT scanning, MRI or bone scanning were requested only if there was a clinical indication. When recurrence was suspected or proven, patients were routinely restaged with bone scans, brain imaging, etc. as appropriate. Information on the disease status of patients not followed up at PMCC was obtained by mail or personal contact with the relevant oncologist or family physician. No patient was lost to follow-up. The date of local or distant progression was taken as the earliest date at which disease progression was confirmed, either clinically (e.g. stridor or hemoptysis subsequently confirmed as due to disease progression by imaging or biopsy) or by imaging or biopsy of asymptomatic lesions.

2.6. Statistical analysis The data have been analysed using a close-out date of 25 July 2002, the earliest of the dates of last contact for the surviving patients. Thus, the status of all patients was known at this date. Survival has been measured from the date of the restaging PET scan to the date of death from any cause. Patients who were still alive at the close-out date had their survival censored at that date. Time to first distant metastasis has been measured from the date of the restaging PET scan to the date of first distant metastasis. Patients who were distant metastasisfree at the earlier of the date of death and the close-out date had their time to distant metastasis censored at that date. The two patients who were assessed as being stage IV on their staging

98

M.P. Mac Manus et al.

PET scans were excluded from the analysis of this endpoint. Time to progression has been measured from the date of the restaging PET scan to the earlier of the date of first locoregional progression and the date of first distant metastasis. Patients who were progression-free at the earlier of the date of death and the close-out date had their time to progression censored at that date. Patients who were assessed as having progressive metabolic disease at the time of the restaging PET or who were known to have progressed prior to the restaging PET scan were excluded from the analysis of this endpoint. Survival estimates were obtained using the Kaplan—Meier method and compared using the Cox proportional hazards regression model. Cumulative incidences were estimated using the method of Kalbfleisch and Prentice [8]. The prognostic significance of individual factors has been summarised using hazard ratios (HR) representing the relative failure rates (death, distant metastasis or progression) for a given group relative to a baseline group. Tests for trend for a given factor were made by assigning the numbers 1, 2, 3, etc. to sequential categories, and including the resulting variable as a linear term in the regression model. The significance of an individual factor in a model was determined by the change in the log-likelihood of the model if that factor were removed from the model. In the multifactor analyses adjusting for pre-treatment ECOG performance status, weight loss and PET stage, the performance status and

Table 1

PET stage (stage after PET scanning) were modelled as linear terms. Weight loss was treated as a binary variable (none versus some). The full PET stages, including ‘‘A’’ and ‘‘B’’ categories (IA, IB, IIA, IIB, IIIA, IIIB and IV), were used, having been assigned the numeric codes 1 through 7, respectively. Ninety-five percent confidence intervals (95% CI) have been given for the main results. No adjustment has been made for multiple comparisons.

3. Results 3.1. Patient characteristics and responses The characteristics of the patients in the various response groups are given in Table 1. Most patients (n = 62, 70%) had stage III disease and few had weight loss > 10% of body mass (n = 4, 5%) or ECOG status worse than 1 (n = 12, 14%). Platinum-based chemotherapy was given concurrently with radiotherapy in 73 patients. Standard chemotherapy during the study period (and given to 54 patients) was single agent carboplatin [9]. Other patients received cisplatin/taxol (n = 7) [10] or cisplatin/5fluorouracil (n = 11) in phase I—II clinical trials. One patient received cisplatin and VP-16. Overall, 40 (45%) of the 88 patients were assessed as achieving CMR on PET, 32 (36%) PMR, 5 (6%) SMD and 11 (13%) PMD. The PET response assessments took place a median of 70 days (2.3 months) after the completion of radiotherapy. Patients who died too soon to

Patient characteristics prior to commencing radiotherapy and PET responses

Factor

Group

CMR

PMR

SMD

PMD

ECOG performance status

0 1 2

6 30 4

6 22 4

0 4 1

2 6 3

14 62 12

Total

Weight loss

none <10% >10%

30 7 3

25 6 1

4 1 0

7 4 0

66 18 4

Clinical stage post—PET

IA IB IIA IIB IIIA IIIB IV

5 2 2 5 9 17 0

0 1 2 3 13 12 1

0 1 0 0 3 1 0

2 1 0 0 4 3 1

7 5 4 8 29 33 2

Histology

Adenocarcinoma Large cell Squamous Mixed NSCLC

7 6 27 0

8 5 19 0

3 0 1 1

5 1 4 1

23 12 51 2

Patterns of failure after PET in NSCLS

99

be assessed or who progressed before their assessment was due, were not included in the study population. Thus, response rates cannot be calculated from the data provided. For this reason, the percentages of patients in each of the response groups have not been included in Table 1.

3.2. Overall survival The duration of potential follow-up from the date of the restaging PET scan to the close-out date ranged from 22 to 63 months with a median of 37.5 months. An estimated 68% (se 5%) of patients survived 1 year after their restaging PET scan, 44% (se 5%) survived 2 years, 28% (se 5%) survived 3 years and 14% (se 6%) survived 4 years (Table 2a, Fig. 1(a)). The estimated median survival was 22 months. Excluding two patients who had stage IV disease on the staging PET scan, an estimated 69% (se 5%) of patients survived 1 year after their restaging PET scan, 44% (se 5%) survived 2 years, 27% (se 5%) survived 3 years and 16% (se 6%) survived 4 years (Table 2a, Fig. 1(a)). The estimated median survival for these patients was 22 months after the restaging PET scan. The most appropriate subgroup for radical RT comprises patients without > 10% weight loss, without PET stage IV disease and with ECOG status 0 or 1. There were 71 such patients in this study. An estimated 73% (se 5%) of these patients survived 1 year after their restaging PET scan, 50% (se 6%) survived 2 years, 32% (se 6%) survived 3 years and 19% (se 7%) survived 4 years. The estimated median survival for these patients was 25 months after the follow up PET scan, which was performed at a median of at least 4 months after the start of treatment.

Table 2a

3.3. Prognostic significance of response assessments A unifactor analysis of the key prognostic factors is given in Table 2b. The prognostic significance of the on-study factors represents the association of these factors with survival following the restaging PET scan and not their association with survival following the initial PET scan. As discussed above, patients who progressed or died before having a response assessment are not included in the study group. Based on a previous study, it is estimated that fewer than 10% of PET-staged patients will have had progressive disease and/or died before their scheduled restaging PET scan [11]. There was a highly significant increase in the death rates associated with worsening response as assessed by PET (Table 2b, Fig. 1(b)). Similar results were obtained when the data were analysed adjusting for ECOG performance status, weight loss and PET stage at the time of commencement of treatment (Table 2c). Most patients in the study had little or no weight loss and had good ECOG performance status; this is reflected in the apparent lack of correlation between either weight loss or ECOG status and overall survival on unifactor analysis.

3.4. Time to first distant metastasis The time to diagnosis of first distant metastasis was studied in patients who were considered not to have distant metastases at the commencement of radiotherapy. Thus, the two patients who were assessed as being stage IV on PET were excluded. Seven (8%) of the remaining 86 patients had distant metastases at the time of the response assessment. The estimated median time to diagnosis of the first distant

Survival estimates n

Estimated median (months)

Estimated % surviving 1 year

Estimated % surviving 2 years

Med.

95% CI

%

95% CI

%

95% CI

All patients

88

22

14—30

68

58—77

44

34—54

By PET response CMR PMR SMD PMD SMD/PMD CMR Not CMR

40 32 5 11 16 40 48

30 13 13 3 5 30 11

23—46 9—25 4—n.a. 2—>35 2—23 23—46 7—17

93 56 60 18 31 90 48

79—98 39—72 20—90 5—51 14—57 80—100 34—62

61 34 20 18 19 58 29

45—75 20—52 3—69 (5—51 6—45 37—90 18—43

n.a.: not assessable.

100

M.P. Mac Manus et al.

Fig. 1 (a) Overall survival and (b) Overall survival by PET scan response.

metastasis was 45 months. An estimated 75% (se 5%) of patients were distant metastasis-free 1 year after their follow—up PET scan, 64% (se 6%) at 2 years, 56% (se 7%) at 3 years and 45% (se 11%) at 4 years (Table 3a, Fig. 2(a)). Patients who were assessed as CMR in the thorax on PET imaging had a significantly decreased risk of

distant metastases relative to those not assessed as CMR (Table 3b, Fig. 2(b)).

3.5. Time to progression Time to progression was studied in patients who were progression-free at the time of the restag-

Patterns of failure after PET in NSCLS

Table 2b

101

Unifactor analysis of survival following follow—up PET scan

Factor

Group

Number of cases

HR

95% CI

p-Value

ECOG performance status

0 1 2

14 62 12

1.00 0.76 2.32

0.39—1.5 0.98—5.5

0.070a

None <10% >10% None Some

66 18 4 66 22

1.00 0.68 1.17 1.00 1.69

0.21—2.2 0.34—4.1

0.063a

0.96—2.9

0.039b

Adenocarcinoma Large cell Squamous Mixed NSCLC

23 12 51 2

1.00 1.50 1.42 1.64

0.20—11 0.18—11 0.23—12

0.94

IA IB IIA IIB IIIA IIIB IV

7 5 4 8 29 33 2

1.00 0.48 0.63 0.27 0.89 0.72 0.72

0.13—1.7 0.16—2.6 0.07—1.1 0.36—2.2 0.29—1.8 0.14—3.6

0.37a

CMR PMR SMD PMD CMR PMR SMD/PMD CMR Not CMR

40 32 5 11 40 32 16 40 48

1.00 2.06 2.63 4.47 1.00 2.06 3.56 1.00 2.40

1.17—3.6 0.99—7.0 2.03—9.8

0.0001a

1.17—3.6 1.81—7.0

0.0001a

1.43—4.0

0.0004b

Weight loss

Histology

Clinical stage post—PET

PET response

a b

One-sided test for trend. One-sided test.

ing PET scan. Thus, 13 patients who were assessed as having progressive metabolic disease (PMD) at the time of the restaging PET scan or who were known to have had stage IV disease (two patients) prior to the restaging PET scan were excluded from the analysis of this endpoint. The estimated median time to progression was 18 months. An estimated 62% (se 65%) of patients

were progression-free 1 year after their follow—up PET scan, 44% (se 6%) at 2 years, 26% (se 7%) at 3 years and 20% (se 8%) at 4 years (Table 4a, Fig. 3(a)). There was a highly significant increase in the rates of progression associated with worsening response as assessed by PET (Table 4b, Fig. 3(b)) both on unifactor analysis and mul-

Table 2c Multifactor analysis of prognostic significance of PET response, adjusting for pre-treatment ECOG performance status, weight loss and PET stage Factor

Group

Number of cases

HR

PET response

CMR PMR SMD/PMD

40 32 16

1.00 2.29 4.06

95%CI

p-Value

1.27—4.1 2.00—8.3

<0.0001

CMR not CMR

48 40

1.00 2.71

1.58—4.7

0.0001

102

M.P. Mac Manus et al.

Fig. 2 (a) Distant-metastasis-free survival and (b) Distant-metastasis-free survival by metabolic response.

tifactor analysis adjusting for ECOG performance status, weight loss and PET stage at the time of commencement of treatment. The cumulative incidence of sites of first progression is given in Table 4c and illustrated in Fig. 4(a)—(c).

4. Discussion The results of this study confirm our early reports that FDG PET response (or as it is becoming more widely known, ‘‘metabolic response’’) assessment after radical RT or chemoRT provides powerful

Patterns of failure after PET in NSCLS

Table 3a

103

Freedom from distant metastasis n

Estimated median time to distant metastasis (months) Med. 95% CI

Estimated % distantmetastasis-free at 1 year % 95% CI

Estimated % distantmetastasis-free at 2 years % 95% CI

All patients

86

45

26—>56

75

64—83

64

52—74

By PET response CMR Not CMR

40 46

>56 29

27—>56 10—>42

87 64

73—95 48—77

77 50

60—89) (34—66)

Excludes 2 PET-assessed stage IV patients.

Table 3b

Association of PET response with time to first distant metastasis

Analysis

Group

Number of casesa

HR

Unifactor

CMR not CMR CMR not CMR

40 46 40 46

1.00 2.69 1.00 2.53

Multifactorb a b

95% CI

p-Value

1.25—5.8

0.0044

1.17—5.5

0.0075

Excludes 2 PET-assessed stage IV patients. Adjusting for pre-treatment ECOG performance status, weight loss and PET stage.

prognostic information. With longer follow up and a larger cohort of patients, it is clear that the differences in survival between the different response categories are maintained over time. Additionally, this study provides new information on the patterns of failure in the different response categories. Our Table 4a

results provide further support for the use of PET to determine treatment response after RT and are consistent with most reports from other groups who have studied the value of PET after RT [12—14] or chemotherapy [15—17]. However, Port and colleagues have reported that PET performed poorly in

Freedom from progression n

Estimated median time to progression (months)

Estimated % progression-free at 1 year

Estimated % progression-free at 2 years

Med.

95% CI

%

(95% CI)

%

(95% CI)

75

18

13—27

62

50—72

44

33—56

By PET response CMR 40 PMR 31 SMD 4

25 10 12

16—>52 4—26 2—>39

75 47 50

59—86 30—65 12—88

54 34 25

38—70 19—54 3—76

All patients

Excludes 13 patients with progressive disease at the time of their restaging PET scan.

Table 4b

Association of PET response with freedom from progression

Analysis

Group

Number of casesa

HR

95% CI

p-Value

Unifactor

CMR PMR SMD

40 31 4

1.00 2.15 2.80

1.17—4.0 0.95—8.2

0.0042

CMR PMR SMD

40 31 4

1.00 2.04 2.49

1.10—3.8 0.84—7.4

0.0086

Multifactor

a b

b

Excludes 13 patients with progressive disease at the time of their restaging PET scan. Adjusting for pre-treatment ECOG performance status, weight loss and PET stage.

104

M.P. Mac Manus et al.

Fig. 3 (a) Time to progression and (b) Time to progression by PET response.

the detection of major pathologic response in primary lung tumors after neoadjuvant chemotherapy. Their study had only 25 patients, used SUV as the only measure of PET response, used histopathologic response rather than survival as their endpoint and assessed PET response at a much earlier time point than was used in our own study. The apparent value

of PET may therefore vary significantly according to study design. Despite excellent survival and freedom from progression in the first 2 years after treatment, patients in the CMR category of our study have shown a disappointingly high rate of relapse and death in the third and fourth years. It is significant that lo-

Patterns of failure after PET in NSCLS cal or locoregional disease failure has contributed to death in the majority of these cases, with relatively few CMR patients dying from distant metastases without local failure. It is likely that some patients who relapsed locally, having attained a CMR,

105 had come very close to ‘‘cure’’ of their lung cancers and could have actually been cured if locoregional therapy had been more effective. The rate of distant metastasis overall was very significantly lower in patients who attained a CMR

Fig. 4 (a) Cumulative incidence of progression, (b) cumulative incidence of progression for CMR patients and (c) cumulative incidence of progression for PMR patients.

106

M.P. Mac Manus et al.

Fig. 4 (Continued).

compared to those who did not. This is a biologically interesting phenomenon. Attainment of an excellent regression of locoregional disease, which for the first time is accurately measurable using PET, is associated with an apparent systemic effect in NSCLC patients treated with radical RT. It is possible that all patients who are destined to experi-

Table 4c

ence distant metastases after RT have already got micrometastatic disease at the time of their primary treatment and that those who are fortunate enough to attain a complete metabolic response simply have tumors with a lower metastatic potential. If that is true, then biological determinants of tumor cell radiosensitivity and metastatic po-

Cumulative incidence of sites of first progression

Year following restaging PET

Locoregional progression only (%)

Locoregional and distant progression (%)

Distant progression only (%)

Total progressions (%)

All patientsa 1 2 3 4

18 25 39 39

(4)b (5) (7) (7)

18 21 26 32

(5) (5) (6) (8)

3 9 9 9

(2) (3) (3) (3)

38 56 74 80

(6) (6) (7) (8)

PET CMR patients 1 10 (5) 2 21 (7) 3 33 (10) 4 33 (10)

13 19 24 33

(5) (7) (8) (11)

3 5 5 5

(3) (4) (4) (4)

25 46 62 71

(7) (8) (10) (11)

PET PMR patients 1 26 (8) 2 31 (9) 3 42 (10)

23 (8) 23 (8) 29 (9)

a b

3 (3) 11 (6) 11 (6)

Excludes 13 patients with progressive disease at the time of their restaging PET scan. Estimated cumulative incidence with estimated standard error in brackets.

53 (9) 66 (9) 83 (8)

Patterns of failure after PET in NSCLS tential are powerfully linked and more intensive local treatments may have little impact on survival. However, it is possible that in many cases treated with radical RT, distant metastasis has not already occurred by the start of treatment and that an effective local therapy can prevent metastasis. This hypothesis is supported by the results of the CHART randomized trial reported by Saunders and colleagues [18], which showed a significantly lower rate of distant metastasis in the more locally effective CHART arm than in the conventional radiotherapy group. The CHART study included no systemic therapy that could potentially have affected microscopic metastatic disease. Patients in the CHART arm also had superior response rates as measured by serial CT scanning compared to the conventional radiotherapy group. An alternative hypothesis to explain our own results is that patients who attained a CR had more chemoresponsive tumors and that systemic therapy delayed or prevented progression with distant metastasis. Patients who attained only a partial response (PMR) to RT in our series had correspondingly higher rates of both locoregional failure and distant metastasis compared to those who attained a CMR. Nevertheless, isolated locoregional failure without distant metastasis was common in these patients, suggesting that had their locoregional therapy been more effective, their survival might also have been longer. The high rate of locoregional failure in patients in this series are a reminder of the importance of locoregional disease control in patients who undergo radical RT after a careful selection process. Few patients had no change in their disease status (stable metabolic disease) on their posttreatment PET scans and those unfortunate patients with progressive disease (PMD) had a median survival of only 3 months compared to 30 months for CMR patients. The patients in this series represent a highly selected group compared to previously-reported series of patients treated with RT for NSCLC. They had undergone staging with FDG—PET prior to commencing treatment. At our center, approximately one third of candidates for radical RT are rejected from this form of therapy because of advanced disease seen on PET [11]. More than 20% of patients with apparent stage III disease have PET-detected distant metastasis [19]. Additionally, those patients who actually received RT were more likely to have had adequate coverage of their tumors within the high dose volume. In about one fourth of patients treated with radical RT after PET, more extensive intra-thoracic disease was apparent on PET than CT scans and adjustment of RT fields to cover these abnormalities almost certainly reduced the risk of

107 a geographic miss [11]. Therefore, in this study, the problem of failure in the thorax is more likely to be a consequence of unfavorable tumor biology and inadequate prescribed radiation dose rather than of technically inadequate treatment. It seems clear that PET can provide significant prognostic information at a relatively early time point after completion of radical RT. How can this new information be used? The first chance for administering aggressive therapy represents the best chance for cure and the results reported here support the use of more intensive locoregional therapy than 60Gy of external beam radiotherapy and concurrent platinum-based chemotherapy as initial treatment. Nevertheless, it may be possible to design trials of therapeutic interventions based on the degree of response to primary therapy as determined by FDG—PET. It is clear that those patients already with progressive disease soon after radical chemoradiation for NSCLC will not benefit significantly from further aggressive therapy, but selected patients with a complete or partial response could potentially have superior freedom from progression or even survival if further response-adapted therapy could be delivered. Such therapy could take the form of surgery, further (perhaps more highly conformal) radiotherapy, chemotherapy, a biological therapy or some combination of more than one of these. There is a suggestion that trimodality therapy (RT plus chemotherapy plus surgery) could [20] produce superior freedom from treatment failure in resectable stage IIIA NSCLC compared to chemoradiation alone, although some thoracotomies in this setting prove to be futile because of unresectable residual tumor. Patients who had undergone futile thoracotomies would certainly have been better served by definitive chemoradiation rather than the less intensive neo-adjuvant therapy. PET could potentially be used to select patients who could benefit from surgery for potentially resectable residual disease after definitive chemoradiation for initially-unresectable disease, although it is not yet clear that this is a feasible treatment strategy. In conclusion, post-treatment PET scans provide accurate prognostic information not available from any other investigation after radiotherapy for NSCLC. Metabolic response stratifies patients into groups with widely differing probabilities of survival, distant metastasis and overall disease progression. Local progression in the absence of distant metastasis is a significant mode of failure for patients who attained a CMR or PMR, suggesting that intensification of locoregional therapy may improve survival in unresectable NSCLC. Prognostic information from PET may assist in the design of future clin-

108

M.P. Mac Manus et al.

ical trials and at the very least will be of value to patients who wish to know their prognosis.

References [1] McLoud TC, Bourgouin PM, Greenberg RW, Kosiuk JP, Templeton PA, Shepard JA, et al. Bronchogenic carcinoma: analysis of staging in the mediastinum with CT by correlative lymph node mapping and sampling. Radiology 1992;182(2):319—23. [2] Lever AM, Henderson D, Ellis DA, Corris PA, Gilmartin JJ. Radiation fibrosis mimicking local recurrence in small cell carcinoma of the bronchus. Br J Radiol 1984;57(674):178— 80. [3] Werner-Wasik M, Xiao Y, Pequignot E, Curran WJ, Hauck W. Assessment of lung cancer response after nonoperative therapy: tumor diameter, bidimensional product, and volume. A serial CT scan-based study. Int J Radiat Oncol Biol Phys 2001;51(1):56—61. [4] Davis SD, Yankelevitz DF, Henschke CI. Radiation effects on the lung: clinical features, pathology, and imaging findings. AJR Am J Roentgenol 1992;159(6):1157—64. [5] Mac Manus MP, Hicks RJ, Matthews JP, McKenzie A, Rischin D, Salminen EK, et al. Positron emission tomography is superior to computed tomography scanning for responseassessment after radical radiotherapy or chemoradiotherapy in patients with non-small-cell lung cancer. J Clin Oncol 2003;21(7):1285—92. [6] Green S, Weiss GR. Southwest Oncology Group standard response criteria, endpoint definitions and toxicity criteria. Invest New Drugs 1992;10(4):239—53. [7] Mountain CF. The new International Staging System for Lung Cancer. Surg Clin North Am 1987;67(5):925—35. [8] Kalbfleisch JD, Prentice RL. The statistical analysis of failure time data. New York, NY: Wiley; 1980. [9] Grossi M, Millward M, Fisher R, Porceddu S, Mac Manus M, Ryan G, et al. Combined modality treatment using concurrent radiotherapy and pharmacologically-guided carboplatin for inoperable and incompletely resected non-small cell lung cancer. Lung Cancer 2001;31(1):73—82. [10] Solomon B, Ball DL, Richardson G, Smith JG, Millward M, MacManus M, et al. Phase I/II study of concurrent twiceweekly paclitaxel and weekly cisplatin with radiation therapy for stage III non-small cell lung cancer. Lung Cancer 2003;41(3):353—61. [11] Mac Manus MP, Hicks RJ, Ball DL, Kalff V, Matthews JP, Salminen E, et al. F-18 fluorodeoxyglucose positron emis-

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

sion tomography staging in radical radiotherapy candidates with nonsmall cell lung carcinoma: powerful correlation with survival and high impact on treatment. Cancer 2001;92(4):886—95. Choi NC, Fischman AJ, Niemierko A, Ryu JS, Lynch T, Wain J, et al. Dose-response relationship between probability of pathologic tumor control and glucose metabolic rate measured with FDG PET after preoperative chemoradiotherapy in locally advanced non-small-cell lung cancer. Int J Radiat Oncol Biol Phys 2002;54(4):1024—35. Erdi YE, Macapinlac H, Rosenzweig KE, Humm JL, Larson SM, Erdi AK, et al. Use of PET to monitor the response of lung cancer to radiation treatment. Eur J Nucl Med 2000;27(7):861—6. Hebert ME, Lowe VJ, Hoffman JM, Patz EF, Anscher MS. Positron emission tomography in the pretreatment evaluation and follow-up of non-small cell lung cancer patients treated with radiotherapy: preliminary findings. Am J Clin Oncol 1996;19(4):416—21. Vansteenkiste JF, Stroobants SG, De Leyn PR, Dupont PJ, Verbeken EK. Potential use of FDG-PET scan after induction chemotherapy in surgically staged IIIa-N2 non-smallcell lung cancer: a prospective pilot study. The Leuven Lung Cancer Group. Ann Oncol 1998;9(11):1193—8. Weber WA, Petersen V, Schmidt B, Tyndale-Hines L, Link T, Peschel C, et al. Positron emission tomography in non-smallcell lung cancer: prediction of response to chemotherapy by quantitative assessment of glucose use. J Clin Oncol 2003;21(14):2651—7. Cerfolio RJ, Ojha B, Mukherjee S, Pask AH, Bass CS, Katholi CR. Positron emission tomography scanning with 2-fluoro-2deoxy-d-glucose as a predictor of response of neoadjuvant treatment for non-small cell carcinoma. J Thorac Cardiovasc Surg 2003;125(4):938—44. Saunders M, Dische S, Barrett A, Harvey A, Gibson D, Parmar M. Continuous hyperfractionated accelerated radiotherapy (CHART) versus conventional radiotherapy in non-small-cell lung cancer: a randomised multicentre trial. CHART Steering Committee. Lancet 1997;350(9072):161—5. MacManus MP, Hicks RJ, Matthews JP, Hogg A, McKenzie AF, Wirth A, et al. High rate of detection of unsuspected distant metastases by pet in apparent stage III non-small-cell lung cancer: implications for radical radiation therapy. Int J Radiat Oncol Biol Phys 2001;50(2):287—93. Albain KS, Crowley JJ, Turrisi III AT, Gandara DR, Farrar WB, Clark JI, et al. Concurrent cisplatin, etoposide, and chest radiotherapy in pathologic stage IIIB non-small-cell lung cancer: a Southwest Oncology Group phase II study, SWOG 9019. J Clin Oncol 2002;20(16):3454—60.