Keywords
recurrence - head and neck cancer - follow-up - salvage
Introduction
The Head and Neck Cancer Guidelines were revised in 2018. This revised edition includes
an additional statement regarding posttreatment follow-up. Here, we report on the
review of conventional follow-up observation after initial treatment at our hospital
based on recurrent cases and investigation on future follow-up observation policies.
Materials and Methods
The study included patients who underwent surgery or chemoradiotherapy (CRT) as initial
treatment during the 5-year period from 2012 to 2016 at the Department of Otorhinolaryngology,
Shinshu University School of Medicine, Matsumoto, Nagano, Japan.
As follow-up observation, patients were examined once per month for the first 2 years
and every 2–3 months thereafter. Those who were treated with chemoradiotherapy (CRT)
underwent computed tomography (CT) (or magnetic resonance imaging [MRI]) in the first
2 months and positron emission tomography (PET) in the first 3 months. For patients
who underwent surgery, baseline images were obtained within 6 months after surgery.
In all patients, imaging examinations were performed every 6 months until 2 years
after the treatment and every 6 to 12 months from the third year onward. Blood sampling
and chest X-rays were also performed once every 6 months.
The diagnosis of recurrence was defined as recurrence being confirmed by pathological
diagnosis or imaging. The duration until recurrence was considered to start from the
treatment start date. For statistical analyses, Mann–Whitney test, Fisher's exact
test, and the log-rank test were performed using PASW statistics version 18 (SPSS
Inc, Chicago, Illinois, United States), and a p-value of less than 0.05 was considered to be statistically significant.
Results
The number of subjects treated via surgery and CRT were 91 and 96, respectively. The
breakdown of subjects in the surgery and CRT groups is shown in [Table 1]. The number of patients with recurrence in the surgery and CRT groups were 20 (22.0%)
and 25 (26.0%), respectively. Three patients in whom clear residual tumors were recognized
after CRT were excluded from the patients with recurrence in this study.
Table 1
Breakdown of the patients included in this study
|
Patients treated with surgery
|
Patients treated with CRT
|
|
Number of patients (total)
|
91
|
96
|
|
Follow-up period, median (mo)
|
37
|
40
|
|
Number of patients with recurrence
|
20
|
25
|
|
Sex (M:F)
|
14:6
|
18:7
|
|
Mean age (y)
|
67.8 (48–85)
|
62.1 (37–76)
|
|
Primary location
|
|
Oral cavity
|
8
|
0
|
|
Epipharynx
|
0
|
5
|
|
Oropharynx
|
1
|
7
|
|
Hypopharynx
|
3
|
8
|
|
Larynx
|
3
|
2
|
|
Nose/paranasal sinus
|
2
|
2
|
|
Other
|
3
|
1
|
|
Stage
|
|
1
|
4
|
0
|
|
2
|
4
|
2
|
|
3
|
2
|
2
|
|
4
|
10
|
21
|
Abbreviations: CRT, chemoradiotherapy; F, female; M, male.
Notes: Oral cancer was common among patients with postsurgery recurrence, and oropharyngeal/hypopharyngeal
cancers were more common in patients with post-CRT recurrence. There were many patients
with advanced cancers among patients with post-CRT recurrences.
Review on Time until Recurrence and Site of Recurrence
Recurrence occurred within the first 2 years in 41 (91.1%) of the 45 patients. The
median interval until recurrence was 12 months, and the mean interval was 12.9 ± 8.3
months. The timing of recurrence in subjects in the surgery and CRT groups has been
shown separately ([Fig. 1]). In the surgery and CRT groups, the mean interval until recurrence was 11.4 ± 9.0
months and 14.2 ± 7.7 months, respectively. No statistical significant difference
was found between the two groups (p = 0.10).
Fig. 1 Interval until recurrence. (A) Patients treated with surgery (n = 20); (B) patients treated with CRT (n = 25). A peak in the number of recurrences was observed at 6–12 months after treatment
in both groups. CRT, chemoradiotherapy.
The site of recurrence was local in 11 patients (24.4%) and cervical lymph nodes in
17 (37.8%), and there was occurrence of distant metastasis in 17 patients (37.8%).
The most frequent site of distant metastasis was the lung (11 patients), followed
by the bone (4 patients), other (2 patients), intra-abdominal lymph nodes (2 patients),
and liver (1 patients). Simultaneous metastases were detected in the bone and intra-abdominal
lymph nodes in one patient.
With respect to each treatment, recurrence was local or in cervical lymph nodes in
approximately 60% of both the surgery and CRT groups. However, cervical recurrence
was seen in 11 patients (55%) in the surgery group, and there were more patients with
local recurrence (9 patients) than with cervical recurrence (6 patients) in the CRT
group ([Fig. 2]).
Fig. 2 The percentage of recurrence sites in patients treated with surgery and CRT. Neck
recurrences were common among the patients treated with surgery, whereas there were
slightly more local recurrences than those of neck recurrences in the patients treated
with CRT. Remote metastasis was seen in ∼40% of both types of patients. CRT, chemoradiotherapy;
LNs, lymph nodes.
Review on Diagnostic Modalities that Led to Diagnosis of Recurrence
The occasions that led to recurrence detection were reviewed mainly by dividing them
into the following categories for analysis: (1) subjective symptoms; (2) findings
at a regular examination (including endoscopy and palpation findings as well as neck
ultrasound examination); and (3) imaging tests (performed as part of regular examination,
such as CT and PET). Patients who complained of symptoms at the time of regular examination
were included in the subjective symptoms category.
Of all subjects with local recurrences, recurrences in 72.7% (8/11) were detected
by either (1) or (2); recurrence in cervical lymph nodes were detected at the same
rate of 47%; (1) + (2) versus (3); and all distant recurrences, 94.1% (16/17), in
subjects were detected by (3), except in one patient ([Fig. 3]).
Fig. 3 Occasions that led to the diagnosis of recurrence. When examining based on recurrent
site, local recurrence was often detected during a medical examination, whereas more
than half of cervical lymph node metastases were detected by diagnostic imaging. Only
one patient with remote metastasis was detected based on subjective symptoms, and
the rest were identified with diagnostic imaging. LNs, lymph nodes.
Secondary Treatment and Prognosis after Diagnosis of Recurrence
[Figs. 4], [5], and [6] present the results of our investigation of secondary treatment and prognosis after
diagnosis of recurrence by recurrence site. Regarding prognosis, deaths in the cancer-bearing
state were included in deaths by primary disease as a rule. However, the death of
one subject owing to another type of cancer was considered as death from another cause.
Fig. 4 Secondary treatment after diagnosis of local recurrence and its prognosis. Surgery
was selected as the secondary treatment in 5/11 (45.5%) patients, prognosis for this
group was relatively good; patients ultimately classified as disease-free survival
accounted for 4/11 (36.4%). BSC, best supportive care.
Fig. 5 Secondary treatment after cervical lymph node recurrence identification and prognosis.
Surgery (neck dissection) was performed in 11/17 (64.7%). Eventually, 5/11 (29.4%)
patients were salvageable. BSC, best supportive care; LN, lymph node.
Fig. 6 Secondary treatment after cervical remote metastasis identification and its prognosis.
Chemotherapy was preferred in majority of the patients. Surgery was performed for
two patients in whom single lung metastasis was identified. One patient maintained
disease-free status. BSC, best supportive care.
Of the patients with local recurrence, salvage surgery was performed in five subjects
(46%) after detection of recurrence, and chemo/re-radiation therapy or best supportive
care was preferred in other patients owing to the difficulties or denial of surgery
([Fig. 4]). With respect to subjects with cervical lymph node recurrence, cervical dissection
was performed in 11 subjects (65%), of which 6 (54.5%) had no recurrence thereafter
([Fig. 5]).
Chemotherapy was preferred for a majority of subjects with distant metastasis, except
two, for whom surgery was performed. Both of these subjects had lung metastasis, and
one exhibited no recurrence after surgery. Disease-free survival was not achieved
in any subject who received chemo/re-radiation therapy (chemo/re-RT; [Fig. 6]).
Finally, the results of the examination of the prognoses of subjects in whom recurrence
occurred by a type of occasion that lead to diagnosis are shown in [Fig. 7]. The median survival time was 2.5 months for the subjective symptom group, 27 months
for the examination group, and 22 months for the imaging group. Thus, subjects in
whom a subjective symptom led to the diagnosis had a poor prognosis (log-rank test,
p < 0.01).
Fig. 7 Survival curve after diagnosis of recurrence according to the type of diagnostic
modalities that led to diagnosis. When setting a censoring event as death owing to
the primary disease, the median survival time for each group from when recurrence
was diagnosed was 2.5 months for the subjective symptom group, 27 months for the examination
group, and 22 months for the diagnostic imaging group. A significant difference was
observed according to the log-rank test (p < 0.01).
Discussion
Although follow-up after treatment of patients with head and neck cancer aims for
early detection and treatment of recurrence, there is no unified opinion on the follow-up
regimen currently. This fact is also stated in the 2018 Head and Neck Cancer Clinical
Practice Guidelines.[1]
The National Comprehensive Cancer Network (NCCN) guidelines state the following schedule
as the follow-up recommendation: (1) medical examination: once every 1 to 3 months
within 1 year after treatment, once every 2 to 6 months in the second year, and once
every 4 to 8 months thereafter; (2) imaging: baseline imaging within 6 months after
treatment; and (3) other: measurement of the level of the thyroid hormone every 6
to 12 months in subjects who have undergone neck irradiation and consideration of
measurement of EBV DNA in subjects with nasopharyngeal cancer.[2] Currently, there is no mention of periodic imaging tests in these guidelines.
The follow-up observation regimen at our hospital is more thorough than that described
in the above-mentioned guidelines in terms of the frequency of medical examination
and, particularly, diagnostic imaging after baseline imaging. [Table 2] compares the Japanese and NCCN guidelines.
Table 2
Comparison of the Japanese Clinical Practice Guidelines for Head and Neck Cancer,
the NCCN Guidelines, and the follow-up observation policy at our hospital
|
Clinical Practice Guidelines for Head and Neck Cancer (2018 edition)
|
NCCN Guidelines (version 2. 2018)
|
Follow-up observation at our hospital
|
|
Medical examination interval
|
No description
|
• ∼Year 1: Once every 1–3 mo
• ∼Year 2: Once every 2–6 mo
• Year 3–5: Once every 4–8 mo
|
• ∼Year 2: every month
• Year 3–5: once every 2–3 mo
|
|
Diagnostic imaging
|
Baseline imaging (3–6 mo after treatment)
• PET is useful for treatment effect evaluation of CRT
|
• Baseline imaging within 6 mo after treatment
• Images should be taken when symptoms/examination findings suggestive of recurrence
are observed
• Chest CT with smoking history
• Regular imaging if lesions are in areas where direct examination is difficult
|
• Baseline imaging: 3–6 mo after surgery
In the patients treated with CRT, CT 2 mo and PET 3 mo after treatment
• Imaging once 6 mo until the end of year 2 after baseline imaging (CT/MRI or PET)
• Once every 6–12 mo in year 3–5
• Chest X-ray: once every 6 mo
|
|
Blood test
|
• Tumor marker: Usefulness not yet confirmed
• Check hypothyroidism
|
• TSH: Once every 6–12 mo
• EBV DNA: only for epipharyngeal cancer
|
• Every 6 mo: blood count, biochemistry, and tumor markers
|
Abbreviations: CRT, chemoradiotherapy; CT, computed tomography; EBV DNA, Epstein-Barr
virus DNA; MRI, magnetic resonance imaging; NCCN, the National Comprehensive Cancer
Network; PET, positron emission tomography; TSH, thyroid-stimulating hormone.
Notes: There is no clear statement about the examination interval in the Clinical
Practice Guidelines for Head and Neck Cancer. Only our follow-up observation policy
indicates that regular imaging tests should be performed after baseline imaging.
Although it has previously been reported that thorough follow-up observation did not
improve the prognosis,[3] high-level follow-up observation is permitted from the viewpoint of quality-adjusted
life years.[4] Thus, this is still debatable. Examination of subjects at our hospital indicated
the possibility that recurrence diagnosis by medical examination and diagnostic imaging
may contribute to improved prognosis. Therefore, although frequent examination and
periodic imaging (including the initial imaging) may be useful, further investigation
is required to determine whether such a thorough follow-up observation improves patients'
quality of life. Our study also showed that blood collection and chest X-ray are not
causes leading to recurrence detection, and previous studies have not confirmed the
utility of these tests either.[5]
[6] Thus, the usage of these tests should be reduced/omitted.
Recurrence occurred within 2 years in 91.1% of patients, with the mean peak at approximately
12 months in our study. The results of the study conducted at another facility by
Kissun et al also reported that in 49 out of 54 patients, recurrence occurred within
2 years after surgery,[7] and this was similar to the results of the present study. Along with the NCCN guidelines,
follow-up observation within the first 2 years was indicated to be important for diagnosis
of recurrence.
Local/cervical lymph node recurrences accounted for approximately 60% in both surgery
and CRT groups. Cervical lymph node recurrences were relatively frequently seen among
the subjects who underwent surgery (11 patients). In 7 of these 11 patients, no cervical
dissection was performed as the initial treatment. Examination of these subjects showed
that three subjects who underwent partial glossectomy and one with hypopharyngeal
cancer who underwent treatment via endoscopic laryngopharyngeal surgery were included.
With respect to tongue cancer, in particular, it has been reported that lymph node
metastasis is present in 30 to 40% of subjects, even when the cancer is at stage I/II
(clinically N0),[8]
[9]
[10] and our study results indicated that careful attention to the diagnosis of cervical
lymph node metastasis is necessary when partial glossectomy without neck dissection
is chosen for the initial treatment.
In this study, remote metastasis was relatively common, and many of these subjects
were diagnosed by conducting regular imaging tests. At our hospital, PET-CT has been
actively performed with local evaluation by conducting CT and MRI in the past, which
may have contributed to the fact that in 12 of the 16 subjects in whom remote metastasis
was detected by imaging, metastasis was actually identified by PET. Except for lung
metastasis, all remote metastases were identified by PET, indicating that PET is extremely
useful in detecting remote metastasis.
PET has been reported to be useful in the diagnosis of recurrence[11]
[12] and is also used in the assessment of objective response to CRT. The noninferiority
of follow-up involving the usage of PET for conventional planned cervical dissection
has been shown[13]; the use of PET in follow-up observation after treatment is important. Conversely,
in the current medical care covered under health insurance, the use of PET for the
diagnosis of recurrence is approved only when recurrence is suspected by conducting
other imaging tests and only when this is observed in other patients. Therefore, for
detecting remote metastasis, cervico-thoraco-abdominal CT appears to be a realistic
policy at this moment.
Investigation of secondary treatment after diagnosis of recurrence and its prognosis
showed that 4/11 (36.4%) subjects with local recurrence, 6/17 (35.3%) with neck recurrence,
and 1/17 (5.9%) with remote metastasis recurrence were classified as having disease-free
status. Because secondary treatment centering on surgery was conducted for almost
all patients with disease-free survival, the main treatment for patients with head
and neck cancer recurrence naturally appeared to be surgery. Therefore, it was considered
important to diagnose local and cervical lymph node recurrence as early as possible
while the metastatic lesions stay in conditions in which surgery is indicated.
Conversely, in reality, there are not many patients for whom surgery can be chosen
as the secondary treatment. Therefore, from this point of view, the survival time
after diagnosis of recurrence and evaluation of the quality of survival also appear
to be important factors. In the patients included in our study, the results of the
investigation of survival time from the diagnosis of recurrence, which was conducted
using the Kaplan–Meier method by considering death resulting from the primary disease
as a censoring event, showed that there was no significant difference between the
surgery and chemo/re-RT groups in terms of secondary treatment (log-rank test; p = 0.77) and that the median survival time was longer in the chemo/re-RT group than
in the surgery group (surgery group: 22 months; chemo/re-RT: 30 months; [Fig. 8]). This result may be influenced by the fact that the chemo/re-RT group includes
cancer-bearing survival to some extent and the follow-up period may be too short to
evaluate the survival time of the different type of secondary treatment.
Fig. 8 Survival curve for each secondary treatment modality after diagnosis of recurrence.
When censoring on death from primary disease, no significant difference was found
in the log-rank test when the surgery and chemo/re-RT treatment groups were compared.
Ope, surgery; Chemo/re-RT, chemotherapy or re-radiation therapy; BSC, best supportive
care.
The analysis of the survival period based on the type of occasion that led to recurrence
diagnosis also showed that the prognosis of the patients, in whom subjective symptoms
led to diagnoses of recurrence, was significantly worse than that of those in which
medical examination or diagnostic imaging led to diagnosis of recurrence ([Fig. 7]). In such analysis, the lead time bias, which is an apparent extended survival period
owing to early detection, cannot be completely excluded. However, 60% (6/10) of patients
with recurrence detected during medical examination remained disease free, whereas
only 14% of those in whom recurrence was detected by subjective symptoms reached the
disease-free survival state after secondary treatment, and this indicated a large
difference between the two. This indicated that detection during a medical examination
may contribute to improved prognosis. In recent years, new drugs, such as nivolumab,[14]
[15] have enabled the extension of patient survival. Thus, it will become even more important
in the future to detect recurrence of the condition in which the quality of life before
appearance of subjective symptoms can be maintained.
The real purpose of follow-up observation may be to improve life prognosis, including
salvage and quality of recurrence cases. Based on this study, issues that need to
be investigated in the future include the following: (1) thorough follow-up observation
focused on patients susceptible to recurrence (in particular, those patients for whom
salvage is either necessary or possible); (2) investigation of treatment methods and
prognosis of the patients in whom recurrence is believed to be able to be detected
early during examination or with the usage of diagnostic imaging techniques; (3) quality
of life during the survival period and analysis, including the economic point of view.
We believe that a prospective comparative study on the earlier-mentioned items will
lead to a truly appropriate follow-up observation regimen.
Summary
We investigated patients in whom surgery or chemoradiotherapy was performed as the
initial treatment and recurrence occurred after that. Because recurrence occurred
in 91.1% of the patients within the first 2 years, this period was considered to be
particularly important. Of the subjects who underwent surgery, particular attention
should be paid to subjects with cervical lymph node metastasis of oral cancer, in
whom neck dissection was not performed. Many distant metastases were detected via
diagnostic imaging, especially by PET, indicating the usefulness of PET for the detection
of metastasis even before the symptoms appear.
In the secondary treatment for recurrence, the main approach used currently that aims
at achieving disease-free survival is the surgical treatment. The diagnosis of recurrence
at a stage when surgery is possible is particularly important for patients in whom
local or cervical lymph node recurrences are involved. In the analysis of prognosis
based on the occasion, prognosis of patients in whom recurrence was diagnosed during
an examination or by imaging was better than that of patients in whom subjective symptoms
led to diagnosis, which was indicative of the possibility that prognosis can be improved
by implementing thorough follow-up observations.
