Survival of histologically proven carcinoma of the lung registered in the North West Thames Region, 1975-1979.

Survival of histologically proven carcinoma of the lung registered in the North West Thames Region, 1975-1979

tions are submitted from some 40 hospitals to the regional registry.
Data were abstracted by one of us (R.J.W.) from the OPCS cancer abstract cards held in the registry. On this card information is required for Treated at this hospital and for Histology confirmed. These data were first explicitly requested on the 1976 OPCS cancer abstract card. Prior to 1976 the information Type of growth was requested only in the form Clinical type of growth and Histological type of growth, if known. Histological confirmation of squamous cell carcinoma, oat cell carcinoma and adenocarcinoma could therefore be assessed for the 1975 records but whether or not a 1975 registered patient received treatment could only be assessed for 3 hospitals in the region. Two, Harefield and Mount Vernon, recorded this information for 1975 in the space on the abstract card headed Notes and for Westminster the information was recorded on the reverse of the abstract card copy retained by Westminster. Data abstraction was for males and females for 1975-1977 and additionally for females [1978][1979]. The sex ratio males: females for lung cancer in England and Wales in 1977 (OPCS, 1982) is 3 7: 1 and the additional 2 years of records for females were obtained to increase the sample size for analysis, otherwise a sub-grouping by histology for females would not have been possible. For the 2 years 1976-1977 with full data for all hospitals and both sexes the sex ratio in the region for histologically confirmed registrations was 3-5: 1.
It would obviously have been impossible in a short space of time to write for followup to, and receive a reply from, the appropriate hospital record departments and general practitioners for those of the 3706 registrations who, according to the North-West Thames Regional Cancer Reg-iFtxy records, had not died. However, by tracing each case using the NHS number and the facilities of the National Health Service Central Registry at Southport, follow-up on whether dead or alive could be rapidly obtained. Hence, following record abstraction from the abstract cards in the regional registry, an initial follow-up was obtained from the NHS Central Registry in October 1980 with a final follow-up in February 1982. In addition, all those registrations alive subsequent to NHS Central Registry follow-up were also subjected to a follow-up enquiry to the cancer registration officers at the hospitals of treatment.
For registrations recorded as untreated, additional enquiries were made to the relevant hospitals to confirm that these lung cancer cases were in fact untreated. Also, the untreated cases who were alive at last NHS Central Registry follow-up were later individually followedup through the cancer registration officers. For the 667 untreated cancer registrations only 5 were still alive at last followup.
Survival rate computations were made using the Hewlett Packard 85 desktop microprocessor of the Westminster Hospital Cancer Registry, Mould (1982), and calculated by a life-table method with data grouped into equal time intervals. The relative survival rates were calculated using the Chester Beatty Research Institute (1974) Serial Abridged Life Tables for England and Wales which are also used by OPCS (1980OPCS ( , 1982.
The age distribution of the 3039 registrations treated for carcinoma of the lung is remarkably similar for males and females, with mean ages respectively 64-5 years and 64-2 years. For the 667 untreated registrations the mean ages are some four years older, 68 1 years for both males and females. These mean ages compare with 66-9 years and 66-3 years for the 1977 OPCS cancer incidence data for males and females, England and Wales, but are higher than those quoted for the United States 1961-1973USDHHS (1976 which are 62 years for males and 59 years for females. The mean ages for each sex for the three main histological groups, squamous cell carcinoma, oat cell carcinoma and adenocarcinoma, do not differ by greater than 3-5 years.
The crude survival rates for untreated carcinoma of the lung in males and females are shown in Fig. 1 for 0-12 months. The one-year relative survival rates for males and females are 9' 9 % and 6 1 % compared with the 9-3% and 5.9% crude 1-year rates. Fig. 2 shows the crude survival rates for 0-60 months for treated carcinoma of the lung in males and females. The 5-year relative survival rates for males and females are 10.7% and 8.3% compared with the 8.6% and 7.3% crude 5-year rates. Table I shows the crude and relative 1year and 5-year survival rates for 3039 treated lung cancer registrations, grouped by sex and histology. Some publications use oat carcinoma synonymously with small cell carcinoma, but since for males 238 registrations were specifically recorded as oat cell compared with only 29 as small cell, and for females 196 oat cell compared with 23 small cell, it was decided to retain the original record descriptions and analyse in terms of specified oat cell carcinoma. Other histological groups such as large cell carcinoma and alveolar cell carcinoma were too small in sample size for analysis.
The most extensive published data on untreated cancer is due to Greenwood (1926) and Shimkin (1951) but neither includes data for untreated lung cancer. Fig. 1 is therefore the only available series of published untreated lung cancer survival rates to date. The 3.4% difference in 1-year survival rates between males  1.-Crude survival rates of untreated carcinoma of the lung in males and females. The relative one-year survival rates for males and females are respectively 9-9% and 6-1%. (Note. Crude rates are plotted in Figures 1 and 2 since these can be computed every month or every 3 months, whereas relative rates can only be computed annually since the CBRI (1974) data only provides nPx probabilities for integral numbers of years between one and five.) and females is probably due to an excess of oat cell carcinoma registrations in the female series. One of the recent clinical trials for treatment of small cell carcinoma was that by the Medical Research Council (1979), and it is interesting to note that they quote an 18% 1-year survival rate for patients treated by radiotherapy, of whom 66% were aged between 55 and 70 years. For oat cell carcinoma in the North West Thames region there is a 13-5% 1-year survival,  one-year survival rates for males and females are respectively 30-8% and 25-1%.
Research Council trial. Livingston et al. (1978) quote a higher 1-year rate of 22% for 73 cases treated by radiotherapy and for a series of 479 cases reviewed by Hansen et al. (1980) which combined several groups of small cell carcinoma treated by various chemotherapy or combination radiotherapy regimes, a 21% 18month survival is quoted for patients with regional disease. This compares with 9.4% and 8.4% 18-month crude survivals in females and males for the series of 238 and 196 registrations in the Table. For a disease such as lung cancer, the 5year survival rate is clearly not a good criterion of treatment success and a 1-year rate is more realistic. For example, for both males and females, the improvement in the 1-year survival rate is some 20% when the patients are treated, compared with when they are untreated. Differences between survivals for males and females is not large for any histological sub-group. However, for adenocarcinoma, the 1-year relative rates differ by more than 5%. In this instance, males have a 1-year survival rate 8% higher than for females, although when the 5-year survival rates are considered, survival is 6% higher for females than for males. A similar difference was also reported in the United States (USDHHS, 1976), for adenocarcinoma, 1960-1973, with a 10% 5-year relative survival rate for males and a 14% 5-year relative survival rate for females.
This review refers only to histologically confirmed registrations since these are "good" data in that a verified histology has been recorded in the routine registration system with sufficient accuracy to enable a survival analysis by histological group to be made. Also, the population of the North West Thames region is such that the regional data can be regarded as a reasonably representative sample from England and Wales. The most recently published OPCS (1982) survival statistics for England and Wales are for 1975 and give 20.0% and 17.5% 1-year relative survival rates for males and females, and 7.0% and 6.6% 5-year relative survival rates for males and females. These are lower than the rates for this North West Thames Regional treated lung cancer series, which have already been stated as respectively 30 8%, 25-1 %, 10.7% and 8.3%. However, this is not surprising since the OPCS analysis is for a combined series of treated and untreated registrations, as well as for histologically confirmed and unconfirmed registrations.
In the North West Thames Region, a significant number of lung cancer registrations are untreated and it is probable that this pattern is repeated throughout England and Wales. Analysis of lung cancer registrations for survival patterns should therefore be made on a basis of treated and untreated series of cases, otherwise a combination of all the registrations cannot be assumed to provide the best estimate of the overall results of treatment within a large population such as a region. When data on histological verification are recorded this should be used for survival analysis, rather than be discarded, as at present occurs at OPCS for England and Wales.